All of lore.kernel.org
 help / color / mirror / Atom feed
* [Qemu-devel] [PATCH 0/5] vhost: enable for all targets
@ 2018-11-26 13:20 Paolo Bonzini
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 1/5] vhost-net: move stubs to a separate file Paolo Bonzini
                   ` (5 more replies)
  0 siblings, 6 replies; 16+ messages in thread
From: Paolo Bonzini @ 2018-11-26 13:20 UTC (permalink / raw)
  To: qemu-devel; +Cc: Marc-André Lureau, Michael S. Tsirkin

vhost does not have to be supported only if KVM is present, in fact
vhost-user does not even need any kind of kernel support.  This series
changes this.  The rationale is that, when vhost-user-test.c will be
converted to qgraph, it will be able to test vhost-user support
for virtio-mmio backend even on x86.

Paolo Bonzini (5):
  vhost-net: move stubs to a separate file
  vhost-net-user: add stubs for when no virtio-net device is present
  vhost: restrict Linux dependency to kernel vhost
  vhost-net: compile it on all targets that have virtio-net.
  vhost-net: revamp configure logic

 backends/Makefile.objs     |   5 +--
 configure                  | 101 ++++++++++++++++++++++++++++-----------------
 default-configs/virtio.mak |   4 +-
 hw/net/Makefile.objs       |   4 +-
 hw/net/vhost_net-stub.c    |  95 ++++++++++++++++++++++++++++++++++++++++++
 hw/net/vhost_net.c         |  78 ++--------------------------------
 hw/virtio/Makefile.objs    |   5 ++-
 hw/virtio/vhost-backend.c  |  11 ++++-
 include/exec/poison.h      |   1 -
 net/Makefile.objs          |   4 +-
 net/net.c                  |   2 +-
 net/vhost-user-stub.c      |  23 +++++++++++
 tests/Makefile.include     |   5 +--
 13 files changed, 209 insertions(+), 129 deletions(-)
 create mode 100644 hw/net/vhost_net-stub.c
 create mode 100644 net/vhost-user-stub.c

-- 
1.8.3.1

^ permalink raw reply	[flat|nested] 16+ messages in thread
* [Qemu-devel] [PATCH 1/5] vhost-net: move stubs to a separate file
  2018-11-26 13:20 [Qemu-devel] [PATCH 0/5] vhost: enable for all targets Paolo Bonzini
@ 2018-11-26 13:20 ` Paolo Bonzini
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 2/5] vhost-net-user: add stubs for when no virtio-net device is present Paolo Bonzini
                   ` (4 subsequent siblings)
  5 siblings, 0 replies; 16+ messages in thread
From: Paolo Bonzini @ 2018-11-26 13:20 UTC (permalink / raw)
  To: qemu-devel; +Cc: Marc-André Lureau, Michael S. Tsirkin

There is no reason for CONFIG_VHOST_NET to be specific to a single target;
it is a host feature that can be add to all targets, as long as they support
the virtio-net device.  Currently CONFIG_VHOST_NET depends on CONFIG_KVM,
but ioeventfd support is present in the core memory API and works with
other accelerators as well.

As a first step, move the vhost-net stubs to a separate file.  Later, they
will become conditional on CONFIG_VIRTIO_NET, which is not available in .c
files.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 hw/net/Makefile.objs    |  4 ++-
 hw/net/vhost_net-stub.c | 95 +++++++++++++++++++++++++++++++++++++++++++++++++
 hw/net/vhost_net.c      | 74 --------------------------------------
 3 files changed, 98 insertions(+), 75 deletions(-)
 create mode 100644 hw/net/vhost_net-stub.c

diff --git a/hw/net/Makefile.objs b/hw/net/Makefile.objs
index fa461d4..c2705e6 100644
--- a/hw/net/Makefile.objs
+++ b/hw/net/Makefile.objs
@@ -37,7 +37,9 @@ obj-$(CONFIG_PSERIES) += spapr_llan.o
 obj-$(CONFIG_XILINX_ETHLITE) += xilinx_ethlite.o
 
 obj-$(CONFIG_VIRTIO_NET) += virtio-net.o
-obj-y += vhost_net.o
+obj-$(CONFIG_VHOST_NET) += vhost_net.o
+common-obj-$(call lnot,$(CONFIG_VHOST_NET)) += vhost_net-stub.o
+common-obj-$(CONFIG_ALL) += vhost_net-stub.o
 
 obj-$(CONFIG_ETSEC) += fsl_etsec/etsec.o fsl_etsec/registers.o \
 			fsl_etsec/rings.o fsl_etsec/miim.o
diff --git a/hw/net/vhost_net-stub.c b/hw/net/vhost_net-stub.c
new file mode 100644
index 0000000..4de1dfb
--- /dev/null
+++ b/hw/net/vhost_net-stub.c
@@ -0,0 +1,95 @@
+/*
+ * vhost-net support
+ *
+ * Copyright Red Hat, Inc. 2010
+ *
+ * Authors:
+ *  Michael S. Tsirkin <mst@redhat.com>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2.  See
+ * the COPYING file in the top-level directory.
+ *
+ * Contributions after 2012-01-13 are licensed under the terms of the
+ * GNU GPL, version 2 or (at your option) any later version.
+ */
+
+#include "qemu/osdep.h"
+#include "net/net.h"
+#include "net/tap.h"
+#include "net/vhost-user.h"
+
+#include "hw/virtio/virtio-net.h"
+#include "net/vhost_net.h"
+#include "qemu/error-report.h"
+
+
+uint64_t vhost_net_get_max_queues(VHostNetState *net)
+{
+    return 1;
+}
+
+struct vhost_net *vhost_net_init(VhostNetOptions *options)
+{
+    error_report("vhost-net support is not compiled in");
+    return NULL;
+}
+
+int vhost_net_start(VirtIODevice *dev,
+                    NetClientState *ncs,
+                    int total_queues)
+{
+    return -ENOSYS;
+}
+void vhost_net_stop(VirtIODevice *dev,
+                    NetClientState *ncs,
+                    int total_queues)
+{
+}
+
+void vhost_net_cleanup(struct vhost_net *net)
+{
+}
+
+uint64_t vhost_net_get_features(struct vhost_net *net, uint64_t features)
+{
+    return features;
+}
+
+void vhost_net_ack_features(struct vhost_net *net, uint64_t features)
+{
+}
+
+uint64_t vhost_net_get_acked_features(VHostNetState *net)
+{
+    return 0;
+}
+
+bool vhost_net_virtqueue_pending(VHostNetState *net, int idx)
+{
+    return false;
+}
+
+void vhost_net_virtqueue_mask(VHostNetState *net, VirtIODevice *dev,
+                              int idx, bool mask)
+{
+}
+
+int vhost_net_notify_migration_done(struct vhost_net *net, char* mac_addr)
+{
+    return -1;
+}
+
+VHostNetState *get_vhost_net(NetClientState *nc)
+{
+    return 0;
+}
+
+int vhost_set_vring_enable(NetClientState *nc, int enable)
+{
+    return 0;
+}
+
+int vhost_net_set_mtu(struct vhost_net *net, uint16_t mtu)
+{
+    return 0;
+}
diff --git a/hw/net/vhost_net.c b/hw/net/vhost_net.c
index e037db6..b901306 100644
--- a/hw/net/vhost_net.c
+++ b/hw/net/vhost_net.c
@@ -23,7 +23,6 @@
 #include "qemu/error-report.h"
 
 
-#ifdef CONFIG_VHOST_NET
 #include <linux/vhost.h>
 #include <sys/socket.h>
 #include <linux/kvm.h>
@@ -449,76 +448,3 @@ int vhost_net_set_mtu(struct vhost_net *net, uint16_t mtu)
 
     return vhost_ops->vhost_net_set_mtu(&net->dev, mtu);
 }
-
-#else
-uint64_t vhost_net_get_max_queues(VHostNetState *net)
-{
-    return 1;
-}
-
-struct vhost_net *vhost_net_init(VhostNetOptions *options)
-{
-    error_report("vhost-net support is not compiled in");
-    return NULL;
-}
-
-int vhost_net_start(VirtIODevice *dev,
-                    NetClientState *ncs,
-                    int total_queues)
-{
-    return -ENOSYS;
-}
-void vhost_net_stop(VirtIODevice *dev,
-                    NetClientState *ncs,
-                    int total_queues)
-{
-}
-
-void vhost_net_cleanup(struct vhost_net *net)
-{
-}
-
-uint64_t vhost_net_get_features(struct vhost_net *net, uint64_t features)
-{
-    return features;
-}
-
-void vhost_net_ack_features(struct vhost_net *net, uint64_t features)
-{
-}
-
-uint64_t vhost_net_get_acked_features(VHostNetState *net)
-{
-    return 0;
-}
-
-bool vhost_net_virtqueue_pending(VHostNetState *net, int idx)
-{
-    return false;
-}
-
-void vhost_net_virtqueue_mask(VHostNetState *net, VirtIODevice *dev,
-                              int idx, bool mask)
-{
-}
-
-int vhost_net_notify_migration_done(struct vhost_net *net, char* mac_addr)
-{
-    return -1;
-}
-
-VHostNetState *get_vhost_net(NetClientState *nc)
-{
-    return 0;
-}
-
-int vhost_set_vring_enable(NetClientState *nc, int enable)
-{
-    return 0;
-}
-
-int vhost_net_set_mtu(struct vhost_net *net, uint16_t mtu)
-{
-    return 0;
-}
-#endif
-- 
1.8.3.1

^ permalink raw reply related	[flat|nested] 16+ messages in thread
* [Qemu-devel] [PATCH 2/5] vhost-net-user: add stubs for when no virtio-net device is present
  2018-11-26 13:20 [Qemu-devel] [PATCH 0/5] vhost: enable for all targets Paolo Bonzini
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 1/5] vhost-net: move stubs to a separate file Paolo Bonzini
@ 2018-11-26 13:20 ` Paolo Bonzini
  2018-11-26 16:45   ` Philippe Mathieu-Daudé
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 3/5] vhost: restrict Linux dependency to kernel vhost Paolo Bonzini
                   ` (3 subsequent siblings)
  5 siblings, 1 reply; 16+ messages in thread
From: Paolo Bonzini @ 2018-11-26 13:20 UTC (permalink / raw)
  To: qemu-devel; +Cc: Marc-André Lureau, Michael S. Tsirkin

hw/net/vhost_net.c needs functions that are declared in net/vhost-user.c: the
vhost-user code is always compiled into QEMU, only the constructor
net_init_vhost_user is unreachable.  Also, net/vhost-user.c needs functions
declared in hw/virtio/vhost-stub.c even if no virtio device exists.

Break this dependency.  First, add a minimal version of net/vhost-user.c,
with no functionality and no dependency on vhost code.  Second, #ifdef out
the calls back to net/vhost-user.c from hw/net/vhost_net.c.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 configure             |  2 +-
 hw/net/vhost_net.c    |  4 ++++
 net/Makefile.objs     |  4 +++-
 net/net.c             |  2 +-
 net/vhost-user-stub.c | 23 +++++++++++++++++++++++
 5 files changed, 32 insertions(+), 3 deletions(-)
 create mode 100644 net/vhost-user-stub.c

diff --git a/configure b/configure
index 0a3c6a7..cda17ef 100755
--- a/configure
+++ b/configure
@@ -6513,7 +6513,7 @@ if test "$vhost_scsi" = "yes" ; then
   echo "CONFIG_VHOST_SCSI=y" >> $config_host_mak
 fi
 if test "$vhost_net" = "yes" -a "$vhost_user" = "yes"; then
-  echo "CONFIG_VHOST_NET_USED=y" >> $config_host_mak
+  echo "CONFIG_VHOST_NET_USER=y" >> $config_host_mak
 fi
 if test "$vhost_crypto" = "yes" ; then
   echo "CONFIG_VHOST_CRYPTO=y" >> $config_host_mak
diff --git a/hw/net/vhost_net.c b/hw/net/vhost_net.c
index b901306..fe6202a 100644
--- a/hw/net/vhost_net.c
+++ b/hw/net/vhost_net.c
@@ -193,6 +193,7 @@ struct vhost_net *vhost_net_init(VhostNetOptions *options)
     }
 
     /* Set sane init value. Override when guest acks. */
+#ifdef CONFIG_VHOST_USER
     if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
         features = vhost_user_get_acked_features(net->nc);
         if (~net->dev.features & features) {
@@ -202,6 +203,7 @@ struct vhost_net *vhost_net_init(VhostNetOptions *options)
             goto fail;
         }
     }
+#endif
 
     vhost_net_ack_features(net, features);
 
@@ -413,10 +415,12 @@ VHostNetState *get_vhost_net(NetClientState *nc)
     case NET_CLIENT_DRIVER_TAP:
         vhost_net = tap_get_vhost_net(nc);
         break;
+#ifdef CONFIG_VHOST_NET_USER
     case NET_CLIENT_DRIVER_VHOST_USER:
         vhost_net = vhost_user_get_vhost_net(nc);
         assert(vhost_net);
         break;
+#endif
     default:
         break;
     }
diff --git a/net/Makefile.objs b/net/Makefile.objs
index b2bf88a..df2b409 100644
--- a/net/Makefile.objs
+++ b/net/Makefile.objs
@@ -3,7 +3,9 @@ common-obj-y += socket.o
 common-obj-y += dump.o
 common-obj-y += eth.o
 common-obj-$(CONFIG_L2TPV3) += l2tpv3.o
-common-obj-$(CONFIG_POSIX) += vhost-user.o
+common-obj-$(call land,$(CONFIG_VIRTIO_NET),$(CONFIG_VHOST_NET_USER)) += vhost-user.o
+common-obj-$(call land,$(call lnot,$(CONFIG_VIRTIO_NET)),$(CONFIG_VHOST_NET_USER)) += vhost-user-stub.o
+common-obj-$(CONFIG_ALL) += vhost-user-stub.o
 common-obj-$(CONFIG_SLIRP) += slirp.o
 common-obj-$(CONFIG_VDE) += vde.o
 common-obj-$(CONFIG_NETMAP) += netmap.o
diff --git a/net/net.c b/net/net.c
index 07c194a..95a74ad 100644
--- a/net/net.c
+++ b/net/net.c
@@ -955,7 +955,7 @@ static int (* const net_client_init_fun[NET_CLIENT_DRIVER__MAX])(
         [NET_CLIENT_DRIVER_BRIDGE]    = net_init_bridge,
 #endif
         [NET_CLIENT_DRIVER_HUBPORT]   = net_init_hubport,
-#ifdef CONFIG_VHOST_NET_USED
+#ifdef CONFIG_VHOST_NET_USER
         [NET_CLIENT_DRIVER_VHOST_USER] = net_init_vhost_user,
 #endif
 #ifdef CONFIG_L2TPV3
diff --git a/net/vhost-user-stub.c b/net/vhost-user-stub.c
new file mode 100644
index 0000000..52ab4e1
--- /dev/null
+++ b/net/vhost-user-stub.c
@@ -0,0 +1,23 @@
+/*
+ * vhost-user-stub.c
+ *
+ * Copyright (c) 2018 Red Hat, Inc.
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
+ *
+ */
+
+#include "qemu/osdep.h"
+#include "clients.h"
+#include "net/vhost_net.h"
+#include "net/vhost-user.h"
+#include "qemu/error-report.h"
+#include "qapi/error.h"
+
+int net_init_vhost_user(const Netdev *netdev, const char *name,
+                        NetClientState *peer, Error **errp)
+{
+    error_setg(errp, "vhost-user requires frontend driver virtio-net-*");
+    return -1;
+}
-- 
1.8.3.1

^ permalink raw reply related	[flat|nested] 16+ messages in thread
* [Qemu-devel] [PATCH 3/5] vhost: restrict Linux dependency to kernel vhost
  2018-11-26 13:20 [Qemu-devel] [PATCH 0/5] vhost: enable for all targets Paolo Bonzini
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 1/5] vhost-net: move stubs to a separate file Paolo Bonzini
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 2/5] vhost-net-user: add stubs for when no virtio-net device is present Paolo Bonzini
@ 2018-11-26 13:20 ` Paolo Bonzini
  2018-11-26 16:51   ` Philippe Mathieu-Daudé
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 4/5] vhost-net: compile it on all targets that have virtio-net Paolo Bonzini
                   ` (2 subsequent siblings)
  5 siblings, 1 reply; 16+ messages in thread
From: Paolo Bonzini @ 2018-11-26 13:20 UTC (permalink / raw)
  To: qemu-devel; +Cc: Marc-André Lureau, Michael S. Tsirkin

vhost-user does not depend on Linux; it can run on any POSIX system.  Restrict
vhost-kernel to Linux in hw/virtio/vhost-backend.c, everything else can be
compiled on all POSIX systems.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 backends/Makefile.objs     |  5 ++---
 default-configs/virtio.mak |  4 ++--
 hw/net/vhost_net.c         |  2 +-
 hw/virtio/Makefile.objs    |  5 +++--
 hw/virtio/vhost-backend.c  | 11 +++++++++--
 5 files changed, 17 insertions(+), 10 deletions(-)

diff --git a/backends/Makefile.objs b/backends/Makefile.objs
index 717fcbd..ff619d3 100644
--- a/backends/Makefile.objs
+++ b/backends/Makefile.objs
@@ -9,10 +9,9 @@ common-obj-$(CONFIG_POSIX) += hostmem-file.o
 common-obj-y += cryptodev.o
 common-obj-y += cryptodev-builtin.o
 
-ifeq ($(CONFIG_VIRTIO),y)
+ifeq ($(CONFIG_VIRTIO_CRYPTO),y)
 common-obj-y += cryptodev-vhost.o
-common-obj-$(call land,$(CONFIG_VHOST_USER),$(CONFIG_LINUX)) += \
-    cryptodev-vhost-user.o
+common-obj-$(CONFIG_VHOST_CRYPTO) += cryptodev-vhost-user.o
 endif
 
 common-obj-$(CONFIG_LINUX) += hostmem-memfd.o
diff --git a/default-configs/virtio.mak b/default-configs/virtio.mak
index 1304849..340050a 100644
--- a/default-configs/virtio.mak
+++ b/default-configs/virtio.mak
@@ -1,5 +1,5 @@
-CONFIG_VHOST_USER_SCSI=$(call land,$(CONFIG_VHOST_USER),$(CONFIG_LINUX))
-CONFIG_VHOST_USER_BLK=$(call land,$(CONFIG_VHOST_USER),$(CONFIG_LINUX))
+CONFIG_VHOST_USER_SCSI=$(CONFIG_VHOST_USER)
+CONFIG_VHOST_USER_BLK=$(CONFIG_VHOST_USER)
 CONFIG_VIRTIO=y
 CONFIG_VIRTIO_9P=y
 CONFIG_VIRTIO_BALLOON=y
diff --git a/hw/net/vhost_net.c b/hw/net/vhost_net.c
index fe6202a..2a300ee 100644
--- a/hw/net/vhost_net.c
+++ b/hw/net/vhost_net.c
@@ -193,7 +193,7 @@ struct vhost_net *vhost_net_init(VhostNetOptions *options)
     }
 
     /* Set sane init value. Override when guest acks. */
-#ifdef CONFIG_VHOST_USER
+#ifdef CONFIG_VHOST_NET_USER
     if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
         features = vhost_user_get_acked_features(net->nc);
         if (~net->dev.features & features) {
diff --git a/hw/virtio/Makefile.objs b/hw/virtio/Makefile.objs
index 1b2799c..e8eff80 100644
--- a/hw/virtio/Makefile.objs
+++ b/hw/virtio/Makefile.objs
@@ -9,9 +9,10 @@ obj-$(CONFIG_VIRTIO_BALLOON) += virtio-balloon.o
 obj-$(CONFIG_VIRTIO_CRYPTO) += virtio-crypto.o
 obj-$(call land,$(CONFIG_VIRTIO_CRYPTO),$(CONFIG_VIRTIO_PCI)) += virtio-crypto-pci.o
 
-obj-$(CONFIG_LINUX) += vhost.o vhost-backend.o vhost-user.o
+obj-$(CONFIG_VHOST_USER) += vhost-user.o
 obj-$(CONFIG_VHOST_VSOCK) += vhost-vsock.o
+obj-$(call lor,$(CONFIG_VHOST_USER),$(CONFIG_LINUX)) += vhost.o vhost-backend.o
+common-obj-$(call lnot,$(call lor,$(CONFIG_VHOST_USER),$(CONFIG_LINUX))) += vhost-stub.o
 endif
 
-common-obj-$(call lnot,$(call land,$(CONFIG_VIRTIO),$(CONFIG_LINUX))) += vhost-stub.o
 common-obj-$(CONFIG_ALL) += vhost-stub.o
diff --git a/hw/virtio/vhost-backend.c b/hw/virtio/vhost-backend.c
index 7f09efa..b5d2e30 100644
--- a/hw/virtio/vhost-backend.c
+++ b/hw/virtio/vhost-backend.c
@@ -9,12 +9,14 @@
  */
 
 #include "qemu/osdep.h"
-#include <linux/vhost.h>
-#include <sys/ioctl.h>
 #include "hw/virtio/vhost.h"
 #include "hw/virtio/vhost-backend.h"
 #include "qemu/error-report.h"
 
+#ifdef CONFIG_LINUX
+#include <linux/vhost.h>
+#include <sys/ioctl.h>
+
 static int vhost_kernel_call(struct vhost_dev *dev, unsigned long int request,
                              void *arg)
 {
@@ -265,18 +267,23 @@ static const VhostOps kernel_ops = {
         .vhost_set_iotlb_callback = vhost_kernel_set_iotlb_callback,
         .vhost_send_device_iotlb_msg = vhost_kernel_send_device_iotlb_msg,
 };
+#endif
 
 int vhost_set_backend_type(struct vhost_dev *dev, VhostBackendType backend_type)
 {
     int r = 0;
 
     switch (backend_type) {
+#ifdef CONFIG_LINUX
     case VHOST_BACKEND_TYPE_KERNEL:
         dev->vhost_ops = &kernel_ops;
         break;
+#endif
+#ifdef CONFIG_VHOST_USER
     case VHOST_BACKEND_TYPE_USER:
         dev->vhost_ops = &user_ops;
         break;
+#endif
     default:
         error_report("Unknown vhost backend type");
         r = -1;
-- 
1.8.3.1

^ permalink raw reply related	[flat|nested] 16+ messages in thread
* [Qemu-devel] [PATCH 4/5] vhost-net: compile it on all targets that have virtio-net.
  2018-11-26 13:20 [Qemu-devel] [PATCH 0/5] vhost: enable for all targets Paolo Bonzini
                   ` (2 preceding siblings ...)
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 3/5] vhost: restrict Linux dependency to kernel vhost Paolo Bonzini
@ 2018-11-26 13:20 ` Paolo Bonzini
  2018-11-26 16:55   ` Philippe Mathieu-Daudé
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 5/5] vhost-net: revamp configure logic Paolo Bonzini
  2018-11-26 22:43 ` [Qemu-devel] [PATCH 0/5] vhost: enable for all targets Michael S. Tsirkin
  5 siblings, 1 reply; 16+ messages in thread
From: Paolo Bonzini @ 2018-11-26 13:20 UTC (permalink / raw)
  To: qemu-devel; +Cc: Marc-André Lureau, Michael S. Tsirkin

This shows a preexisting bug: if a KVM target did not have virtio-net enabled,
it would fail with undefined symbols when vhost was enabled.  This must now
be fixed, lest targets that have no virtio-net fail to compile.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 configure              | 11 ++++-------
 hw/net/Makefile.objs   |  4 ++--
 include/exec/poison.h  |  1 -
 tests/Makefile.include |  5 +----
 4 files changed, 7 insertions(+), 14 deletions(-)

diff --git a/configure b/configure
index cda17ef..b3b4464 100755
--- a/configure
+++ b/configure
@@ -6512,7 +6512,10 @@ fi
 if test "$vhost_scsi" = "yes" ; then
   echo "CONFIG_VHOST_SCSI=y" >> $config_host_mak
 fi
-if test "$vhost_net" = "yes" -a "$vhost_user" = "yes"; then
+if test "$vhost_net" = "yes" ; then
+  echo "CONFIG_VHOST_NET=y" >> $config_host_mak
+fi
+if test "$vhost_net_user" = "yes" ; then
   echo "CONFIG_VHOST_NET_USER=y" >> $config_host_mak
 fi
 if test "$vhost_crypto" = "yes" ; then
@@ -7275,12 +7278,6 @@ if supported_xen_target $target; then
 fi
 if supported_kvm_target $target; then
     echo "CONFIG_KVM=y" >> $config_target_mak
-    if test "$vhost_net" = "yes" ; then
-        echo "CONFIG_VHOST_NET=y" >> $config_target_mak
-        if test "$vhost_user" = "yes" ; then
-            echo "CONFIG_VHOST_USER_NET_TEST_$target_name=y" >> $config_host_mak
-        fi
-    fi
 fi
 if supported_hax_target $target; then
     echo "CONFIG_HAX=y" >> $config_target_mak
diff --git a/hw/net/Makefile.objs b/hw/net/Makefile.objs
index c2705e6..2d7ee0a 100644
--- a/hw/net/Makefile.objs
+++ b/hw/net/Makefile.objs
@@ -37,8 +37,8 @@ obj-$(CONFIG_PSERIES) += spapr_llan.o
 obj-$(CONFIG_XILINX_ETHLITE) += xilinx_ethlite.o
 
 obj-$(CONFIG_VIRTIO_NET) += virtio-net.o
-obj-$(CONFIG_VHOST_NET) += vhost_net.o
-common-obj-$(call lnot,$(CONFIG_VHOST_NET)) += vhost_net-stub.o
+common-obj-$(call land,$(CONFIG_VIRTIO_NET),$(CONFIG_VHOST_NET)) += vhost_net.o
+common-obj-$(call lnot,$(call land,$(CONFIG_VIRTIO_NET),$(CONFIG_VHOST_NET))) += vhost_net-stub.o
 common-obj-$(CONFIG_ALL) += vhost_net-stub.o
 
 obj-$(CONFIG_ETSEC) += fsl_etsec/etsec.o fsl_etsec/registers.o \
diff --git a/include/exec/poison.h b/include/exec/poison.h
index 32d5378..b158632 100644
--- a/include/exec/poison.h
+++ b/include/exec/poison.h
@@ -85,7 +85,6 @@
 #pragma GCC poison CONFIG_XTENSA_DIS
 
 #pragma GCC poison CONFIG_LINUX_USER
-#pragma GCC poison CONFIG_VHOST_NET
 #pragma GCC poison CONFIG_KVM
 #pragma GCC poison CONFIG_SOFTMMU
 
diff --git a/tests/Makefile.include b/tests/Makefile.include
index fb0b449..03a64ce 100644
--- a/tests/Makefile.include
+++ b/tests/Makefile.include
@@ -204,10 +204,7 @@ check-qtest-i386-$(CONFIG_USB_XHCI_NEC) += tests/usb-hcd-xhci-test$(EXESUF)
 check-qtest-i386-y += tests/cpu-plug-test$(EXESUF)
 check-qtest-i386-y += tests/q35-test$(EXESUF)
 check-qtest-i386-y += tests/vmgenid-test$(EXESUF)
-check-qtest-i386-$(CONFIG_VHOST_USER_NET_TEST_i386) += tests/vhost-user-test$(EXESUF)
-ifeq ($(CONFIG_VHOST_USER_NET_TEST_i386),)
-check-qtest-x86_64-$(CONFIG_VHOST_USER_NET_TEST_x86_64) += tests/vhost-user-test$(EXESUF)
-endif
+check-qtest-i386-$(CONFIG_VHOST_NET_USER) += tests/vhost-user-test$(EXESUF)
 check-qtest-i386-$(CONFIG_TPM_CRB) += tests/tpm-crb-swtpm-test$(EXESUF)
 check-qtest-i386-$(CONFIG_TPM_CRB) += tests/tpm-crb-test$(EXESUF)
 check-qtest-i386-$(CONFIG_TPM_TIS) += tests/tpm-tis-swtpm-test$(EXESUF)
-- 
1.8.3.1

^ permalink raw reply related	[flat|nested] 16+ messages in thread
* [Qemu-devel] [PATCH 5/5] vhost-net: revamp configure logic
  2018-11-26 13:20 [Qemu-devel] [PATCH 0/5] vhost: enable for all targets Paolo Bonzini
                   ` (3 preceding siblings ...)
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 4/5] vhost-net: compile it on all targets that have virtio-net Paolo Bonzini
@ 2018-11-26 13:20 ` Paolo Bonzini
  2018-11-26 17:00   ` Philippe Mathieu-Daudé
  2018-11-26 22:43 ` [Qemu-devel] [PATCH 0/5] vhost: enable for all targets Michael S. Tsirkin
  5 siblings, 1 reply; 16+ messages in thread
From: Paolo Bonzini @ 2018-11-26 13:20 UTC (permalink / raw)
  To: qemu-devel; +Cc: Marc-André Lureau, Michael S. Tsirkin

Detect all invalid configurations (e.g. mingw32 with vhost-user,
non-Linux with vhost-kernel).  As a collateral benefit, all vhost-kernel
backends can be now disabled if one wants to reduce the attack surface.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 configure                 | 88 +++++++++++++++++++++++++++++++----------------
 hw/virtio/Makefile.objs   |  4 +--
 hw/virtio/vhost-backend.c |  4 +--
 3 files changed, 62 insertions(+), 34 deletions(-)

diff --git a/configure b/configure
index b3b4464..a15f143 100755
--- a/configure
+++ b/configure
@@ -366,10 +366,10 @@ libattr=""
 xfs=""
 tcg="yes"
 membarrier=""
-vhost_net="no"
-vhost_crypto="no"
-vhost_scsi="no"
-vhost_vsock="no"
+vhost_net=""
+vhost_crypto=""
+vhost_scsi=""
+vhost_vsock=""
 vhost_user=""
 kvm="no"
 hax="no"
@@ -774,6 +774,7 @@ case $targetos in
 MINGW32*)
   mingw32="yes"
   hax="yes"
+  vhost_user="no"
   audio_possible_drivers="dsound sdl"
   if check_include dsound.h; then
     audio_drv_list="dsound"
@@ -874,10 +875,6 @@ Linux)
   linux="yes"
   linux_user="yes"
   kvm="yes"
-  vhost_net="yes"
-  vhost_crypto="yes"
-  vhost_scsi="yes"
-  vhost_vsock="yes"
   QEMU_INCLUDES="-I\$(SRC_PATH)/linux-headers -I$PWD/linux-headers $QEMU_INCLUDES"
   supported_os="yes"
   libudev="yes"
@@ -1258,11 +1255,7 @@ for opt do
   ;;
   --disable-vhost-crypto) vhost_crypto="no"
   ;;
-  --enable-vhost-crypto)
-      vhost_crypto="yes"
-      if test "$mingw32" = "yes"; then
-          error_exit "vhost-crypto isn't available on win32"
-      fi
+  --enable-vhost-crypto) vhost_crypto="yes"
   ;;
   --disable-vhost-scsi) vhost_scsi="no"
   ;;
@@ -1463,11 +1456,11 @@ for opt do
   ;;
   --disable-vhost-user) vhost_user="no"
   ;;
-  --enable-vhost-user)
-      vhost_user="yes"
-      if test "$mingw32" = "yes"; then
-          error_exit "vhost-user isn't available on win32"
-      fi
+  --enable-vhost-user) vhost_user="yes"
+  ;;
+  --disable-vhost-kernel) vhost_kernel="no"
+  ;;
+  --enable-vhost-kernel) vhost_kernel="yes"
   ;;
   --disable-capstone) capstone="no"
   ;;
@@ -1499,14 +1492,6 @@ for opt do
   esac
 done
 
-if test "$vhost_user" = ""; then
-    if test "$mingw32" = "yes"; then
-        vhost_user="no"
-    else
-        vhost_user="yes"
-    fi
-fi
-
 case "$cpu" in
     ppc)
            CPU_CFLAGS="-m32"
@@ -1731,8 +1716,12 @@ disabled with --disable-FEATURE, default is enabled if available:
   linux-aio       Linux AIO support
   cap-ng          libcap-ng support
   attr            attr and xattr support
-  vhost-net       vhost-net acceleration support
-  vhost-crypto    vhost-crypto acceleration support
+  vhost-net       vhost-net kernel acceleration support
+  vhost-vsock     virtio sockets device support
+  vhost-scsi      vhost-scsi kernel target support
+  vhost-crypto    vhost-user-crypto backend support
+  vhost-kernel    vhost kernel backend support
+  vhost-user      vhost-user backend support
   spice           spice
   rbd             rados block device (rbd)
   libiscsi        iscsi support
@@ -1756,7 +1744,6 @@ disabled with --disable-FEATURE, default is enabled if available:
   jemalloc        jemalloc support
   avx2            AVX2 optimization support
   replication     replication support
-  vhost-vsock     virtio sockets device support
   opengl          opengl support
   virglrenderer   virgl rendering support
   xfsctl          xfsctl support
@@ -1773,7 +1760,6 @@ disabled with --disable-FEATURE, default is enabled if available:
   parallels       parallels image format support
   sheepdog        sheepdog block driver support
   crypto-afalg    Linux AF_ALG crypto backend driver
-  vhost-user      vhost-user support
   capstone        capstone disassembler support
   debug-mutex     mutex debugging support
   libpmem         libpmem support
@@ -2150,6 +2136,45 @@ else
   l2tpv3=no
 fi
 
+#########################################
+# vhost interdependencies and host support
+
+# vhost backends
+test "$vhost_user" = "" && vhost_user=yes
+if test "$vhost_user" = "yes" && test "$mingw32" = "yes"; then
+  error_exit "vhost-user isn't available on win32"
+fi
+test "$vhost_kernel" = "" && vhost_kernel=$linux
+if test "$vhost_kernel" = "yes" && test "$linux" != "yes"; then
+  error_exit "vhost-kernel is only available on Linux"
+fi
+
+# vhost-kernel devices
+test "$vhost_scsi" = "" && vhost_scsi=$vhost_kernel
+if test "$vhost_scsi" = "yes" && test "$vhost_kernel" != "yes"; then
+  error_exit "--enable-vhost-scsi requires --enable-vhost-kernel"
+fi
+test "$vhost_vsock" = "" && vhost_vsock=$vhost_kernel
+if test "$vhost_vsock" = "yes" && test "$vhost_kernel" != "yes"; then
+  error_exit "--enable-vhost-vsock requires --enable-vhost-kernel"
+fi
+
+# vhost-user backends
+test "$vhost_net_user" = "" && vhost_net_user=$vhost_user
+if test "$vhost_net_user" = "yes" && test "$vhost_user" = "no"; then
+  error_exit "--enable-vhost-net-user requires --enable-vhost-user"
+fi
+test "$vhost_crypto" = "" && vhost_crypto=$vhost_user
+if test "$vhost_crypto" = "yes" && test "$vhost_user" = "no"; then
+  error_exit "--enable-vhost-crypto requires --enable-vhost-user"
+fi
+
+# OR the vhost-kernel and vhost-user values for simplicity
+if test "$vhost_net" = ""; then
+  test "$vhost_net_user" = "yes" && vhost_net=yes
+  test "$vhost_kernel" = "yes" && vhost_net=yes
+fi
+
 ##########################################
 # MinGW / Mingw-w64 localtime_r/gmtime_r check
 
@@ -6524,6 +6549,9 @@ fi
 if test "$vhost_vsock" = "yes" ; then
   echo "CONFIG_VHOST_VSOCK=y" >> $config_host_mak
 fi
+if test "$vhost_kernel" = "yes" ; then
+  echo "CONFIG_VHOST_KERNEL=y" >> $config_host_mak
+fi
 if test "$vhost_user" = "yes" ; then
   echo "CONFIG_VHOST_USER=y" >> $config_host_mak
 fi
diff --git a/hw/virtio/Makefile.objs b/hw/virtio/Makefile.objs
index e8eff80..87402d1 100644
--- a/hw/virtio/Makefile.objs
+++ b/hw/virtio/Makefile.objs
@@ -11,8 +11,8 @@ obj-$(call land,$(CONFIG_VIRTIO_CRYPTO),$(CONFIG_VIRTIO_PCI)) += virtio-crypto-p
 
 obj-$(CONFIG_VHOST_USER) += vhost-user.o
 obj-$(CONFIG_VHOST_VSOCK) += vhost-vsock.o
-obj-$(call lor,$(CONFIG_VHOST_USER),$(CONFIG_LINUX)) += vhost.o vhost-backend.o
-common-obj-$(call lnot,$(call lor,$(CONFIG_VHOST_USER),$(CONFIG_LINUX))) += vhost-stub.o
+obj-$(call lor,$(CONFIG_VHOST_USER),$(CONFIG_VHOST_KERNEL)) += vhost.o vhost-backend.o
+common-obj-$(call lnot,$(call lor,$(CONFIG_VHOST_USER),$(CONFIG_VHOST_KERNEL))) += vhost-stub.o
 endif
 
 common-obj-$(CONFIG_ALL) += vhost-stub.o
diff --git a/hw/virtio/vhost-backend.c b/hw/virtio/vhost-backend.c
index b5d2e30..46d388b 100644
--- a/hw/virtio/vhost-backend.c
+++ b/hw/virtio/vhost-backend.c
@@ -13,7 +13,7 @@
 #include "hw/virtio/vhost-backend.h"
 #include "qemu/error-report.h"
 
-#ifdef CONFIG_LINUX
+#ifdef CONFIG_VHOST_KERNEL
 #include <linux/vhost.h>
 #include <sys/ioctl.h>
 
@@ -274,7 +274,7 @@ int vhost_set_backend_type(struct vhost_dev *dev, VhostBackendType backend_type)
     int r = 0;
 
     switch (backend_type) {
-#ifdef CONFIG_LINUX
+#ifdef CONFIG_VHOST_KERNEL
     case VHOST_BACKEND_TYPE_KERNEL:
         dev->vhost_ops = &kernel_ops;
         break;
-- 
1.8.3.1

^ permalink raw reply related	[flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 2/5] vhost-net-user: add stubs for when no virtio-net device is present
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 2/5] vhost-net-user: add stubs for when no virtio-net device is present Paolo Bonzini
@ 2018-11-26 16:45   ` Philippe Mathieu-Daudé
  2018-11-26 22:43     ` Michael S. Tsirkin
  0 siblings, 1 reply; 16+ messages in thread
From: Philippe Mathieu-Daudé @ 2018-11-26 16:45 UTC (permalink / raw)
  To: Paolo Bonzini, qemu-devel; +Cc: Marc-André Lureau, Michael S. Tsirkin

On 26/11/18 14:20, Paolo Bonzini wrote:
> hw/net/vhost_net.c needs functions that are declared in net/vhost-user.c: the
> vhost-user code is always compiled into QEMU, only the constructor
> net_init_vhost_user is unreachable.  Also, net/vhost-user.c needs functions
> declared in hw/virtio/vhost-stub.c even if no virtio device exists.
> 
> Break this dependency.  First, add a minimal version of net/vhost-user.c,
> with no functionality and no dependency on vhost code.  Second, #ifdef out
> the calls back to net/vhost-user.c from hw/net/vhost_net.c.
> 
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>

> ---
>  configure             |  2 +-
>  hw/net/vhost_net.c    |  4 ++++
>  net/Makefile.objs     |  4 +++-
>  net/net.c             |  2 +-
>  net/vhost-user-stub.c | 23 +++++++++++++++++++++++
>  5 files changed, 32 insertions(+), 3 deletions(-)
>  create mode 100644 net/vhost-user-stub.c
> 
> diff --git a/configure b/configure
> index 0a3c6a7..cda17ef 100755
> --- a/configure
> +++ b/configure
> @@ -6513,7 +6513,7 @@ if test "$vhost_scsi" = "yes" ; then
>    echo "CONFIG_VHOST_SCSI=y" >> $config_host_mak
>  fi
>  if test "$vhost_net" = "yes" -a "$vhost_user" = "yes"; then
> -  echo "CONFIG_VHOST_NET_USED=y" >> $config_host_mak
> +  echo "CONFIG_VHOST_NET_USER=y" >> $config_host_mak
>  fi
>  if test "$vhost_crypto" = "yes" ; then
>    echo "CONFIG_VHOST_CRYPTO=y" >> $config_host_mak
> diff --git a/hw/net/vhost_net.c b/hw/net/vhost_net.c
> index b901306..fe6202a 100644
> --- a/hw/net/vhost_net.c
> +++ b/hw/net/vhost_net.c
> @@ -193,6 +193,7 @@ struct vhost_net *vhost_net_init(VhostNetOptions *options)
>      }
>  
>      /* Set sane init value. Override when guest acks. */
> +#ifdef CONFIG_VHOST_USER
>      if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
>          features = vhost_user_get_acked_features(net->nc);
>          if (~net->dev.features & features) {
> @@ -202,6 +203,7 @@ struct vhost_net *vhost_net_init(VhostNetOptions *options)
>              goto fail;
>          }
>      }
> +#endif
>  
>      vhost_net_ack_features(net, features);
>  
> @@ -413,10 +415,12 @@ VHostNetState *get_vhost_net(NetClientState *nc)
>      case NET_CLIENT_DRIVER_TAP:
>          vhost_net = tap_get_vhost_net(nc);
>          break;
> +#ifdef CONFIG_VHOST_NET_USER
>      case NET_CLIENT_DRIVER_VHOST_USER:
>          vhost_net = vhost_user_get_vhost_net(nc);
>          assert(vhost_net);
>          break;
> +#endif
>      default:
>          break;
>      }
> diff --git a/net/Makefile.objs b/net/Makefile.objs
> index b2bf88a..df2b409 100644
> --- a/net/Makefile.objs
> +++ b/net/Makefile.objs
> @@ -3,7 +3,9 @@ common-obj-y += socket.o
>  common-obj-y += dump.o
>  common-obj-y += eth.o
>  common-obj-$(CONFIG_L2TPV3) += l2tpv3.o
> -common-obj-$(CONFIG_POSIX) += vhost-user.o
> +common-obj-$(call land,$(CONFIG_VIRTIO_NET),$(CONFIG_VHOST_NET_USER)) += vhost-user.o
> +common-obj-$(call land,$(call lnot,$(CONFIG_VIRTIO_NET)),$(CONFIG_VHOST_NET_USER)) += vhost-user-stub.o
> +common-obj-$(CONFIG_ALL) += vhost-user-stub.o
>  common-obj-$(CONFIG_SLIRP) += slirp.o
>  common-obj-$(CONFIG_VDE) += vde.o
>  common-obj-$(CONFIG_NETMAP) += netmap.o
> diff --git a/net/net.c b/net/net.c
> index 07c194a..95a74ad 100644
> --- a/net/net.c
> +++ b/net/net.c
> @@ -955,7 +955,7 @@ static int (* const net_client_init_fun[NET_CLIENT_DRIVER__MAX])(
>          [NET_CLIENT_DRIVER_BRIDGE]    = net_init_bridge,
>  #endif
>          [NET_CLIENT_DRIVER_HUBPORT]   = net_init_hubport,
> -#ifdef CONFIG_VHOST_NET_USED
> +#ifdef CONFIG_VHOST_NET_USER
>          [NET_CLIENT_DRIVER_VHOST_USER] = net_init_vhost_user,
>  #endif
>  #ifdef CONFIG_L2TPV3
> diff --git a/net/vhost-user-stub.c b/net/vhost-user-stub.c
> new file mode 100644
> index 0000000..52ab4e1
> --- /dev/null
> +++ b/net/vhost-user-stub.c
> @@ -0,0 +1,23 @@
> +/*
> + * vhost-user-stub.c
> + *
> + * Copyright (c) 2018 Red Hat, Inc.
> + *
> + * This work is licensed under the terms of the GNU GPL, version 2 or later.
> + * See the COPYING file in the top-level directory.
> + *
> + */
> +
> +#include "qemu/osdep.h"
> +#include "clients.h"
> +#include "net/vhost_net.h"
> +#include "net/vhost-user.h"
> +#include "qemu/error-report.h"
> +#include "qapi/error.h"
> +
> +int net_init_vhost_user(const Netdev *netdev, const char *name,
> +                        NetClientState *peer, Error **errp)
> +{
> +    error_setg(errp, "vhost-user requires frontend driver virtio-net-*");
> +    return -1;
> +}
> 

^ permalink raw reply	[flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 3/5] vhost: restrict Linux dependency to kernel vhost
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 3/5] vhost: restrict Linux dependency to kernel vhost Paolo Bonzini
@ 2018-11-26 16:51   ` Philippe Mathieu-Daudé
  0 siblings, 0 replies; 16+ messages in thread
From: Philippe Mathieu-Daudé @ 2018-11-26 16:51 UTC (permalink / raw)
  To: Paolo Bonzini, qemu-devel; +Cc: Marc-André Lureau, Michael S. Tsirkin

On 26/11/18 14:20, Paolo Bonzini wrote:
> vhost-user does not depend on Linux; it can run on any POSIX system.  Restrict
> vhost-kernel to Linux in hw/virtio/vhost-backend.c, everything else can be
> compiled on all POSIX systems.
> 
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
>  backends/Makefile.objs     |  5 ++---
>  default-configs/virtio.mak |  4 ++--
>  hw/net/vhost_net.c         |  2 +-
>  hw/virtio/Makefile.objs    |  5 +++--
>  hw/virtio/vhost-backend.c  | 11 +++++++++--
>  5 files changed, 17 insertions(+), 10 deletions(-)
> 
> diff --git a/backends/Makefile.objs b/backends/Makefile.objs
> index 717fcbd..ff619d3 100644
> --- a/backends/Makefile.objs
> +++ b/backends/Makefile.objs
> @@ -9,10 +9,9 @@ common-obj-$(CONFIG_POSIX) += hostmem-file.o
>  common-obj-y += cryptodev.o
>  common-obj-y += cryptodev-builtin.o
>  
> -ifeq ($(CONFIG_VIRTIO),y)
> +ifeq ($(CONFIG_VIRTIO_CRYPTO),y)
>  common-obj-y += cryptodev-vhost.o
> -common-obj-$(call land,$(CONFIG_VHOST_USER),$(CONFIG_LINUX)) += \
> -    cryptodev-vhost-user.o
> +common-obj-$(CONFIG_VHOST_CRYPTO) += cryptodev-vhost-user.o
>  endif

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>

>  
>  common-obj-$(CONFIG_LINUX) += hostmem-memfd.o
> diff --git a/default-configs/virtio.mak b/default-configs/virtio.mak
> index 1304849..340050a 100644
> --- a/default-configs/virtio.mak
> +++ b/default-configs/virtio.mak
> @@ -1,5 +1,5 @@
> -CONFIG_VHOST_USER_SCSI=$(call land,$(CONFIG_VHOST_USER),$(CONFIG_LINUX))
> -CONFIG_VHOST_USER_BLK=$(call land,$(CONFIG_VHOST_USER),$(CONFIG_LINUX))
> +CONFIG_VHOST_USER_SCSI=$(CONFIG_VHOST_USER)
> +CONFIG_VHOST_USER_BLK=$(CONFIG_VHOST_USER)
>  CONFIG_VIRTIO=y
>  CONFIG_VIRTIO_9P=y
>  CONFIG_VIRTIO_BALLOON=y
> diff --git a/hw/net/vhost_net.c b/hw/net/vhost_net.c
> index fe6202a..2a300ee 100644
> --- a/hw/net/vhost_net.c
> +++ b/hw/net/vhost_net.c
> @@ -193,7 +193,7 @@ struct vhost_net *vhost_net_init(VhostNetOptions *options)
>      }
>  
>      /* Set sane init value. Override when guest acks. */
> -#ifdef CONFIG_VHOST_USER
> +#ifdef CONFIG_VHOST_NET_USER
>      if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
>          features = vhost_user_get_acked_features(net->nc);
>          if (~net->dev.features & features) {
> diff --git a/hw/virtio/Makefile.objs b/hw/virtio/Makefile.objs
> index 1b2799c..e8eff80 100644
> --- a/hw/virtio/Makefile.objs
> +++ b/hw/virtio/Makefile.objs
> @@ -9,9 +9,10 @@ obj-$(CONFIG_VIRTIO_BALLOON) += virtio-balloon.o
>  obj-$(CONFIG_VIRTIO_CRYPTO) += virtio-crypto.o
>  obj-$(call land,$(CONFIG_VIRTIO_CRYPTO),$(CONFIG_VIRTIO_PCI)) += virtio-crypto-pci.o
>  
> -obj-$(CONFIG_LINUX) += vhost.o vhost-backend.o vhost-user.o
> +obj-$(CONFIG_VHOST_USER) += vhost-user.o
>  obj-$(CONFIG_VHOST_VSOCK) += vhost-vsock.o
> +obj-$(call lor,$(CONFIG_VHOST_USER),$(CONFIG_LINUX)) += vhost.o vhost-backend.o
> +common-obj-$(call lnot,$(call lor,$(CONFIG_VHOST_USER),$(CONFIG_LINUX))) += vhost-stub.o
>  endif
>  
> -common-obj-$(call lnot,$(call land,$(CONFIG_VIRTIO),$(CONFIG_LINUX))) += vhost-stub.o
>  common-obj-$(CONFIG_ALL) += vhost-stub.o
> diff --git a/hw/virtio/vhost-backend.c b/hw/virtio/vhost-backend.c
> index 7f09efa..b5d2e30 100644
> --- a/hw/virtio/vhost-backend.c
> +++ b/hw/virtio/vhost-backend.c
> @@ -9,12 +9,14 @@
>   */
>  
>  #include "qemu/osdep.h"
> -#include <linux/vhost.h>
> -#include <sys/ioctl.h>
>  #include "hw/virtio/vhost.h"
>  #include "hw/virtio/vhost-backend.h"
>  #include "qemu/error-report.h"
>  
> +#ifdef CONFIG_LINUX
> +#include <linux/vhost.h>
> +#include <sys/ioctl.h>
> +
>  static int vhost_kernel_call(struct vhost_dev *dev, unsigned long int request,
>                               void *arg)
>  {
> @@ -265,18 +267,23 @@ static const VhostOps kernel_ops = {
>          .vhost_set_iotlb_callback = vhost_kernel_set_iotlb_callback,
>          .vhost_send_device_iotlb_msg = vhost_kernel_send_device_iotlb_msg,
>  };
> +#endif
>  
>  int vhost_set_backend_type(struct vhost_dev *dev, VhostBackendType backend_type)
>  {
>      int r = 0;
>  
>      switch (backend_type) {
> +#ifdef CONFIG_LINUX
>      case VHOST_BACKEND_TYPE_KERNEL:
>          dev->vhost_ops = &kernel_ops;
>          break;
> +#endif
> +#ifdef CONFIG_VHOST_USER
>      case VHOST_BACKEND_TYPE_USER:
>          dev->vhost_ops = &user_ops;
>          break;
> +#endif
>      default:
>          error_report("Unknown vhost backend type");
>          r = -1;
> 

^ permalink raw reply	[flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 4/5] vhost-net: compile it on all targets that have virtio-net.
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 4/5] vhost-net: compile it on all targets that have virtio-net Paolo Bonzini
@ 2018-11-26 16:55   ` Philippe Mathieu-Daudé
  0 siblings, 0 replies; 16+ messages in thread
From: Philippe Mathieu-Daudé @ 2018-11-26 16:55 UTC (permalink / raw)
  To: Paolo Bonzini, qemu-devel; +Cc: Marc-André Lureau, Michael S. Tsirkin

On 26/11/18 14:20, Paolo Bonzini wrote:
> This shows a preexisting bug: if a KVM target did not have virtio-net enabled,
> it would fail with undefined symbols when vhost was enabled.  This must now
> be fixed, lest targets that have no virtio-net fail to compile.
> 
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>

> ---
>  configure              | 11 ++++-------
>  hw/net/Makefile.objs   |  4 ++--
>  include/exec/poison.h  |  1 -
>  tests/Makefile.include |  5 +----
>  4 files changed, 7 insertions(+), 14 deletions(-)
> 
> diff --git a/configure b/configure
> index cda17ef..b3b4464 100755
> --- a/configure
> +++ b/configure
> @@ -6512,7 +6512,10 @@ fi
>  if test "$vhost_scsi" = "yes" ; then
>    echo "CONFIG_VHOST_SCSI=y" >> $config_host_mak
>  fi
> -if test "$vhost_net" = "yes" -a "$vhost_user" = "yes"; then
> +if test "$vhost_net" = "yes" ; then
> +  echo "CONFIG_VHOST_NET=y" >> $config_host_mak
> +fi
> +if test "$vhost_net_user" = "yes" ; then
>    echo "CONFIG_VHOST_NET_USER=y" >> $config_host_mak
>  fi
>  if test "$vhost_crypto" = "yes" ; then
> @@ -7275,12 +7278,6 @@ if supported_xen_target $target; then
>  fi
>  if supported_kvm_target $target; then
>      echo "CONFIG_KVM=y" >> $config_target_mak
> -    if test "$vhost_net" = "yes" ; then
> -        echo "CONFIG_VHOST_NET=y" >> $config_target_mak
> -        if test "$vhost_user" = "yes" ; then
> -            echo "CONFIG_VHOST_USER_NET_TEST_$target_name=y" >> $config_host_mak
> -        fi
> -    fi
>  fi
>  if supported_hax_target $target; then
>      echo "CONFIG_HAX=y" >> $config_target_mak
> diff --git a/hw/net/Makefile.objs b/hw/net/Makefile.objs
> index c2705e6..2d7ee0a 100644
> --- a/hw/net/Makefile.objs
> +++ b/hw/net/Makefile.objs
> @@ -37,8 +37,8 @@ obj-$(CONFIG_PSERIES) += spapr_llan.o
>  obj-$(CONFIG_XILINX_ETHLITE) += xilinx_ethlite.o
>  
>  obj-$(CONFIG_VIRTIO_NET) += virtio-net.o
> -obj-$(CONFIG_VHOST_NET) += vhost_net.o
> -common-obj-$(call lnot,$(CONFIG_VHOST_NET)) += vhost_net-stub.o
> +common-obj-$(call land,$(CONFIG_VIRTIO_NET),$(CONFIG_VHOST_NET)) += vhost_net.o
> +common-obj-$(call lnot,$(call land,$(CONFIG_VIRTIO_NET),$(CONFIG_VHOST_NET))) += vhost_net-stub.o
>  common-obj-$(CONFIG_ALL) += vhost_net-stub.o
>  
>  obj-$(CONFIG_ETSEC) += fsl_etsec/etsec.o fsl_etsec/registers.o \
> diff --git a/include/exec/poison.h b/include/exec/poison.h
> index 32d5378..b158632 100644
> --- a/include/exec/poison.h
> +++ b/include/exec/poison.h
> @@ -85,7 +85,6 @@
>  #pragma GCC poison CONFIG_XTENSA_DIS
>  
>  #pragma GCC poison CONFIG_LINUX_USER
> -#pragma GCC poison CONFIG_VHOST_NET
>  #pragma GCC poison CONFIG_KVM
>  #pragma GCC poison CONFIG_SOFTMMU
>  
> diff --git a/tests/Makefile.include b/tests/Makefile.include
> index fb0b449..03a64ce 100644
> --- a/tests/Makefile.include
> +++ b/tests/Makefile.include
> @@ -204,10 +204,7 @@ check-qtest-i386-$(CONFIG_USB_XHCI_NEC) += tests/usb-hcd-xhci-test$(EXESUF)
>  check-qtest-i386-y += tests/cpu-plug-test$(EXESUF)
>  check-qtest-i386-y += tests/q35-test$(EXESUF)
>  check-qtest-i386-y += tests/vmgenid-test$(EXESUF)
> -check-qtest-i386-$(CONFIG_VHOST_USER_NET_TEST_i386) += tests/vhost-user-test$(EXESUF)
> -ifeq ($(CONFIG_VHOST_USER_NET_TEST_i386),)
> -check-qtest-x86_64-$(CONFIG_VHOST_USER_NET_TEST_x86_64) += tests/vhost-user-test$(EXESUF)
> -endif
> +check-qtest-i386-$(CONFIG_VHOST_NET_USER) += tests/vhost-user-test$(EXESUF)
>  check-qtest-i386-$(CONFIG_TPM_CRB) += tests/tpm-crb-swtpm-test$(EXESUF)
>  check-qtest-i386-$(CONFIG_TPM_CRB) += tests/tpm-crb-test$(EXESUF)
>  check-qtest-i386-$(CONFIG_TPM_TIS) += tests/tpm-tis-swtpm-test$(EXESUF)
> 

^ permalink raw reply	[flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 5/5] vhost-net: revamp configure logic
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 5/5] vhost-net: revamp configure logic Paolo Bonzini
@ 2018-11-26 17:00   ` Philippe Mathieu-Daudé
  0 siblings, 0 replies; 16+ messages in thread
From: Philippe Mathieu-Daudé @ 2018-11-26 17:00 UTC (permalink / raw)
  To: Paolo Bonzini, qemu-devel; +Cc: Marc-André Lureau, Michael S. Tsirkin

On 26/11/18 14:20, Paolo Bonzini wrote:
> Detect all invalid configurations (e.g. mingw32 with vhost-user,
> non-Linux with vhost-kernel).  As a collateral benefit, all vhost-kernel
> backends can be now disabled if one wants to reduce the attack surface.
> 
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
>  configure                 | 88 +++++++++++++++++++++++++++++++----------------
>  hw/virtio/Makefile.objs   |  4 +--
>  hw/virtio/vhost-backend.c |  4 +--
>  3 files changed, 62 insertions(+), 34 deletions(-)
> 
> diff --git a/configure b/configure
> index b3b4464..a15f143 100755
> --- a/configure
> +++ b/configure
> @@ -366,10 +366,10 @@ libattr=""
>  xfs=""
>  tcg="yes"
>  membarrier=""
> -vhost_net="no"
> -vhost_crypto="no"
> -vhost_scsi="no"
> -vhost_vsock="no"
> +vhost_net=""
> +vhost_crypto=""
> +vhost_scsi=""
> +vhost_vsock=""
>  vhost_user=""
>  kvm="no"
>  hax="no"
> @@ -774,6 +774,7 @@ case $targetos in
>  MINGW32*)
>    mingw32="yes"
>    hax="yes"
> +  vhost_user="no"
>    audio_possible_drivers="dsound sdl"
>    if check_include dsound.h; then
>      audio_drv_list="dsound"
> @@ -874,10 +875,6 @@ Linux)
>    linux="yes"
>    linux_user="yes"
>    kvm="yes"
> -  vhost_net="yes"
> -  vhost_crypto="yes"
> -  vhost_scsi="yes"
> -  vhost_vsock="yes"
>    QEMU_INCLUDES="-I\$(SRC_PATH)/linux-headers -I$PWD/linux-headers $QEMU_INCLUDES"
>    supported_os="yes"
>    libudev="yes"
> @@ -1258,11 +1255,7 @@ for opt do
>    ;;
>    --disable-vhost-crypto) vhost_crypto="no"
>    ;;
> -  --enable-vhost-crypto)
> -      vhost_crypto="yes"
> -      if test "$mingw32" = "yes"; then
> -          error_exit "vhost-crypto isn't available on win32"
> -      fi
> +  --enable-vhost-crypto) vhost_crypto="yes"
>    ;;
>    --disable-vhost-scsi) vhost_scsi="no"
>    ;;
> @@ -1463,11 +1456,11 @@ for opt do
>    ;;
>    --disable-vhost-user) vhost_user="no"
>    ;;
> -  --enable-vhost-user)
> -      vhost_user="yes"
> -      if test "$mingw32" = "yes"; then
> -          error_exit "vhost-user isn't available on win32"
> -      fi
> +  --enable-vhost-user) vhost_user="yes"
> +  ;;
> +  --disable-vhost-kernel) vhost_kernel="no"
> +  ;;
> +  --enable-vhost-kernel) vhost_kernel="yes"
>    ;;
>    --disable-capstone) capstone="no"
>    ;;
> @@ -1499,14 +1492,6 @@ for opt do
>    esac
>  done
>  
> -if test "$vhost_user" = ""; then
> -    if test "$mingw32" = "yes"; then
> -        vhost_user="no"
> -    else
> -        vhost_user="yes"
> -    fi
> -fi
> -
>  case "$cpu" in
>      ppc)
>             CPU_CFLAGS="-m32"
> @@ -1731,8 +1716,12 @@ disabled with --disable-FEATURE, default is enabled if available:
>    linux-aio       Linux AIO support
>    cap-ng          libcap-ng support
>    attr            attr and xattr support
> -  vhost-net       vhost-net acceleration support
> -  vhost-crypto    vhost-crypto acceleration support
> +  vhost-net       vhost-net kernel acceleration support
> +  vhost-vsock     virtio sockets device support
> +  vhost-scsi      vhost-scsi kernel target support
> +  vhost-crypto    vhost-user-crypto backend support
> +  vhost-kernel    vhost kernel backend support
> +  vhost-user      vhost-user backend support
>    spice           spice
>    rbd             rados block device (rbd)
>    libiscsi        iscsi support
> @@ -1756,7 +1744,6 @@ disabled with --disable-FEATURE, default is enabled if available:
>    jemalloc        jemalloc support
>    avx2            AVX2 optimization support
>    replication     replication support
> -  vhost-vsock     virtio sockets device support
>    opengl          opengl support
>    virglrenderer   virgl rendering support
>    xfsctl          xfsctl support
> @@ -1773,7 +1760,6 @@ disabled with --disable-FEATURE, default is enabled if available:
>    parallels       parallels image format support
>    sheepdog        sheepdog block driver support
>    crypto-afalg    Linux AF_ALG crypto backend driver
> -  vhost-user      vhost-user support
>    capstone        capstone disassembler support
>    debug-mutex     mutex debugging support
>    libpmem         libpmem support
> @@ -2150,6 +2136,45 @@ else
>    l2tpv3=no
>  fi
>  
> +#########################################
> +# vhost interdependencies and host support
> +
> +# vhost backends
> +test "$vhost_user" = "" && vhost_user=yes
> +if test "$vhost_user" = "yes" && test "$mingw32" = "yes"; then
> +  error_exit "vhost-user isn't available on win32"
> +fi
> +test "$vhost_kernel" = "" && vhost_kernel=$linux
> +if test "$vhost_kernel" = "yes" && test "$linux" != "yes"; then
> +  error_exit "vhost-kernel is only available on Linux"
> +fi
> +
> +# vhost-kernel devices
> +test "$vhost_scsi" = "" && vhost_scsi=$vhost_kernel
> +if test "$vhost_scsi" = "yes" && test "$vhost_kernel" != "yes"; then
> +  error_exit "--enable-vhost-scsi requires --enable-vhost-kernel"
> +fi
> +test "$vhost_vsock" = "" && vhost_vsock=$vhost_kernel
> +if test "$vhost_vsock" = "yes" && test "$vhost_kernel" != "yes"; then
> +  error_exit "--enable-vhost-vsock requires --enable-vhost-kernel"
> +fi
> +
> +# vhost-user backends
> +test "$vhost_net_user" = "" && vhost_net_user=$vhost_user
> +if test "$vhost_net_user" = "yes" && test "$vhost_user" = "no"; then
> +  error_exit "--enable-vhost-net-user requires --enable-vhost-user"
> +fi
> +test "$vhost_crypto" = "" && vhost_crypto=$vhost_user
> +if test "$vhost_crypto" = "yes" && test "$vhost_user" = "no"; then
> +  error_exit "--enable-vhost-crypto requires --enable-vhost-user"
> +fi
> +
> +# OR the vhost-kernel and vhost-user values for simplicity
> +if test "$vhost_net" = ""; then
> +  test "$vhost_net_user" = "yes" && vhost_net=yes
> +  test "$vhost_kernel" = "yes" && vhost_net=yes
> +fi

Easier to review, thanks.

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>

> +
>  ##########################################
>  # MinGW / Mingw-w64 localtime_r/gmtime_r check
>  
> @@ -6524,6 +6549,9 @@ fi
>  if test "$vhost_vsock" = "yes" ; then
>    echo "CONFIG_VHOST_VSOCK=y" >> $config_host_mak
>  fi
> +if test "$vhost_kernel" = "yes" ; then
> +  echo "CONFIG_VHOST_KERNEL=y" >> $config_host_mak
> +fi
>  if test "$vhost_user" = "yes" ; then
>    echo "CONFIG_VHOST_USER=y" >> $config_host_mak
>  fi
> diff --git a/hw/virtio/Makefile.objs b/hw/virtio/Makefile.objs
> index e8eff80..87402d1 100644
> --- a/hw/virtio/Makefile.objs
> +++ b/hw/virtio/Makefile.objs
> @@ -11,8 +11,8 @@ obj-$(call land,$(CONFIG_VIRTIO_CRYPTO),$(CONFIG_VIRTIO_PCI)) += virtio-crypto-p
>  
>  obj-$(CONFIG_VHOST_USER) += vhost-user.o
>  obj-$(CONFIG_VHOST_VSOCK) += vhost-vsock.o
> -obj-$(call lor,$(CONFIG_VHOST_USER),$(CONFIG_LINUX)) += vhost.o vhost-backend.o
> -common-obj-$(call lnot,$(call lor,$(CONFIG_VHOST_USER),$(CONFIG_LINUX))) += vhost-stub.o
> +obj-$(call lor,$(CONFIG_VHOST_USER),$(CONFIG_VHOST_KERNEL)) += vhost.o vhost-backend.o
> +common-obj-$(call lnot,$(call lor,$(CONFIG_VHOST_USER),$(CONFIG_VHOST_KERNEL))) += vhost-stub.o
>  endif
>  
>  common-obj-$(CONFIG_ALL) += vhost-stub.o
> diff --git a/hw/virtio/vhost-backend.c b/hw/virtio/vhost-backend.c
> index b5d2e30..46d388b 100644
> --- a/hw/virtio/vhost-backend.c
> +++ b/hw/virtio/vhost-backend.c
> @@ -13,7 +13,7 @@
>  #include "hw/virtio/vhost-backend.h"
>  #include "qemu/error-report.h"
>  
> -#ifdef CONFIG_LINUX
> +#ifdef CONFIG_VHOST_KERNEL
>  #include <linux/vhost.h>
>  #include <sys/ioctl.h>
>  
> @@ -274,7 +274,7 @@ int vhost_set_backend_type(struct vhost_dev *dev, VhostBackendType backend_type)
>      int r = 0;
>  
>      switch (backend_type) {
> -#ifdef CONFIG_LINUX
> +#ifdef CONFIG_VHOST_KERNEL
>      case VHOST_BACKEND_TYPE_KERNEL:
>          dev->vhost_ops = &kernel_ops;
>          break;
> 

^ permalink raw reply	[flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 0/5] vhost: enable for all targets
  2018-11-26 13:20 [Qemu-devel] [PATCH 0/5] vhost: enable for all targets Paolo Bonzini
                   ` (4 preceding siblings ...)
  2018-11-26 13:20 ` [Qemu-devel] [PATCH 5/5] vhost-net: revamp configure logic Paolo Bonzini
@ 2018-11-26 22:43 ` Michael S. Tsirkin
  2018-11-27  9:04   ` Paolo Bonzini
  5 siblings, 1 reply; 16+ messages in thread
From: Michael S. Tsirkin @ 2018-11-26 22:43 UTC (permalink / raw)
  To: Paolo Bonzini; +Cc: qemu-devel, Marc-André Lureau

On Mon, Nov 26, 2018 at 02:20:38PM +0100, Paolo Bonzini wrote:
> vhost does not have to be supported only if KVM is present, in fact
> vhost-user does not even need any kind of kernel support.  This series
> changes this.  The rationale is that, when vhost-user-test.c will be
> converted to qgraph, it will be able to test vhost-user support
> for virtio-mmio backend even on x86.

The reason for limiting it to KVM was very simple:
it has the same set of problems with ordering
as mttcg.

So I guess it's fine but I think we must then limit it
to when tcg emits fence instructions.

Otherwise we'll get subtle race conditions.



> Paolo Bonzini (5):
>   vhost-net: move stubs to a separate file
>   vhost-net-user: add stubs for when no virtio-net device is present
>   vhost: restrict Linux dependency to kernel vhost
>   vhost-net: compile it on all targets that have virtio-net.
>   vhost-net: revamp configure logic
> 
>  backends/Makefile.objs     |   5 +--
>  configure                  | 101 ++++++++++++++++++++++++++++-----------------
>  default-configs/virtio.mak |   4 +-
>  hw/net/Makefile.objs       |   4 +-
>  hw/net/vhost_net-stub.c    |  95 ++++++++++++++++++++++++++++++++++++++++++
>  hw/net/vhost_net.c         |  78 ++--------------------------------
>  hw/virtio/Makefile.objs    |   5 ++-
>  hw/virtio/vhost-backend.c  |  11 ++++-
>  include/exec/poison.h      |   1 -
>  net/Makefile.objs          |   4 +-
>  net/net.c                  |   2 +-
>  net/vhost-user-stub.c      |  23 +++++++++++
>  tests/Makefile.include     |   5 +--
>  13 files changed, 209 insertions(+), 129 deletions(-)
>  create mode 100644 hw/net/vhost_net-stub.c
>  create mode 100644 net/vhost-user-stub.c
> 
> -- 
> 1.8.3.1

^ permalink raw reply	[flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 2/5] vhost-net-user: add stubs for when no virtio-net device is present
  2018-11-26 16:45   ` Philippe Mathieu-Daudé
@ 2018-11-26 22:43     ` Michael S. Tsirkin
  2018-11-26 22:52       ` Philippe Mathieu-Daudé
  0 siblings, 1 reply; 16+ messages in thread
From: Michael S. Tsirkin @ 2018-11-26 22:43 UTC (permalink / raw)
  To: Philippe Mathieu-Daudé
  Cc: Paolo Bonzini, qemu-devel, Marc-André Lureau

On Mon, Nov 26, 2018 at 05:45:47PM +0100, Philippe Mathieu-Daudé wrote:
> On 26/11/18 14:20, Paolo Bonzini wrote:
> > hw/net/vhost_net.c needs functions that are declared in net/vhost-user.c: the
> > vhost-user code is always compiled into QEMU, only the constructor
> > net_init_vhost_user is unreachable.  Also, net/vhost-user.c needs functions
> > declared in hw/virtio/vhost-stub.c even if no virtio device exists.
> > 
> > Break this dependency.  First, add a minimal version of net/vhost-user.c,
> > with no functionality and no dependency on vhost code.  Second, #ifdef out
> > the calls back to net/vhost-user.c from hw/net/vhost_net.c.
> > 
> > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> 
> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
> Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>

whic hosts/targets and how did you test exactly?

> > ---
> >  configure             |  2 +-
> >  hw/net/vhost_net.c    |  4 ++++
> >  net/Makefile.objs     |  4 +++-
> >  net/net.c             |  2 +-
> >  net/vhost-user-stub.c | 23 +++++++++++++++++++++++
> >  5 files changed, 32 insertions(+), 3 deletions(-)
> >  create mode 100644 net/vhost-user-stub.c
> > 
> > diff --git a/configure b/configure
> > index 0a3c6a7..cda17ef 100755
> > --- a/configure
> > +++ b/configure
> > @@ -6513,7 +6513,7 @@ if test "$vhost_scsi" = "yes" ; then
> >    echo "CONFIG_VHOST_SCSI=y" >> $config_host_mak
> >  fi
> >  if test "$vhost_net" = "yes" -a "$vhost_user" = "yes"; then
> > -  echo "CONFIG_VHOST_NET_USED=y" >> $config_host_mak
> > +  echo "CONFIG_VHOST_NET_USER=y" >> $config_host_mak
> >  fi
> >  if test "$vhost_crypto" = "yes" ; then
> >    echo "CONFIG_VHOST_CRYPTO=y" >> $config_host_mak
> > diff --git a/hw/net/vhost_net.c b/hw/net/vhost_net.c
> > index b901306..fe6202a 100644
> > --- a/hw/net/vhost_net.c
> > +++ b/hw/net/vhost_net.c
> > @@ -193,6 +193,7 @@ struct vhost_net *vhost_net_init(VhostNetOptions *options)
> >      }
> >  
> >      /* Set sane init value. Override when guest acks. */
> > +#ifdef CONFIG_VHOST_USER
> >      if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
> >          features = vhost_user_get_acked_features(net->nc);
> >          if (~net->dev.features & features) {
> > @@ -202,6 +203,7 @@ struct vhost_net *vhost_net_init(VhostNetOptions *options)
> >              goto fail;
> >          }
> >      }
> > +#endif
> >  
> >      vhost_net_ack_features(net, features);
> >  
> > @@ -413,10 +415,12 @@ VHostNetState *get_vhost_net(NetClientState *nc)
> >      case NET_CLIENT_DRIVER_TAP:
> >          vhost_net = tap_get_vhost_net(nc);
> >          break;
> > +#ifdef CONFIG_VHOST_NET_USER
> >      case NET_CLIENT_DRIVER_VHOST_USER:
> >          vhost_net = vhost_user_get_vhost_net(nc);
> >          assert(vhost_net);
> >          break;
> > +#endif
> >      default:
> >          break;
> >      }
> > diff --git a/net/Makefile.objs b/net/Makefile.objs
> > index b2bf88a..df2b409 100644
> > --- a/net/Makefile.objs
> > +++ b/net/Makefile.objs
> > @@ -3,7 +3,9 @@ common-obj-y += socket.o
> >  common-obj-y += dump.o
> >  common-obj-y += eth.o
> >  common-obj-$(CONFIG_L2TPV3) += l2tpv3.o
> > -common-obj-$(CONFIG_POSIX) += vhost-user.o
> > +common-obj-$(call land,$(CONFIG_VIRTIO_NET),$(CONFIG_VHOST_NET_USER)) += vhost-user.o
> > +common-obj-$(call land,$(call lnot,$(CONFIG_VIRTIO_NET)),$(CONFIG_VHOST_NET_USER)) += vhost-user-stub.o
> > +common-obj-$(CONFIG_ALL) += vhost-user-stub.o
> >  common-obj-$(CONFIG_SLIRP) += slirp.o
> >  common-obj-$(CONFIG_VDE) += vde.o
> >  common-obj-$(CONFIG_NETMAP) += netmap.o
> > diff --git a/net/net.c b/net/net.c
> > index 07c194a..95a74ad 100644
> > --- a/net/net.c
> > +++ b/net/net.c
> > @@ -955,7 +955,7 @@ static int (* const net_client_init_fun[NET_CLIENT_DRIVER__MAX])(
> >          [NET_CLIENT_DRIVER_BRIDGE]    = net_init_bridge,
> >  #endif
> >          [NET_CLIENT_DRIVER_HUBPORT]   = net_init_hubport,
> > -#ifdef CONFIG_VHOST_NET_USED
> > +#ifdef CONFIG_VHOST_NET_USER
> >          [NET_CLIENT_DRIVER_VHOST_USER] = net_init_vhost_user,
> >  #endif
> >  #ifdef CONFIG_L2TPV3
> > diff --git a/net/vhost-user-stub.c b/net/vhost-user-stub.c
> > new file mode 100644
> > index 0000000..52ab4e1
> > --- /dev/null
> > +++ b/net/vhost-user-stub.c
> > @@ -0,0 +1,23 @@
> > +/*
> > + * vhost-user-stub.c
> > + *
> > + * Copyright (c) 2018 Red Hat, Inc.
> > + *
> > + * This work is licensed under the terms of the GNU GPL, version 2 or later.
> > + * See the COPYING file in the top-level directory.
> > + *
> > + */
> > +
> > +#include "qemu/osdep.h"
> > +#include "clients.h"
> > +#include "net/vhost_net.h"
> > +#include "net/vhost-user.h"
> > +#include "qemu/error-report.h"
> > +#include "qapi/error.h"
> > +
> > +int net_init_vhost_user(const Netdev *netdev, const char *name,
> > +                        NetClientState *peer, Error **errp)
> > +{
> > +    error_setg(errp, "vhost-user requires frontend driver virtio-net-*");
> > +    return -1;
> > +}
> > 

^ permalink raw reply	[flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 2/5] vhost-net-user: add stubs for when no virtio-net device is present
  2018-11-26 22:43     ` Michael S. Tsirkin
@ 2018-11-26 22:52       ` Philippe Mathieu-Daudé
  0 siblings, 0 replies; 16+ messages in thread
From: Philippe Mathieu-Daudé @ 2018-11-26 22:52 UTC (permalink / raw)
  To: Michael S. Tsirkin; +Cc: Paolo Bonzini, qemu-devel, Marc-André Lureau

On 26/11/18 23:43, Michael S. Tsirkin wrote:
> On Mon, Nov 26, 2018 at 05:45:47PM +0100, Philippe Mathieu-Daudé wrote:
>> On 26/11/18 14:20, Paolo Bonzini wrote:
>>> hw/net/vhost_net.c needs functions that are declared in net/vhost-user.c: the
>>> vhost-user code is always compiled into QEMU, only the constructor
>>> net_init_vhost_user is unreachable.  Also, net/vhost-user.c needs functions
>>> declared in hw/virtio/vhost-stub.c even if no virtio device exists.
>>>
>>> Break this dependency.  First, add a minimal version of net/vhost-user.c,
>>> with no functionality and no dependency on vhost code.  Second, #ifdef out
>>> the calls back to net/vhost-user.c from hw/net/vhost_net.c.
>>>
>>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>>
>> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
>> Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
> 
> whic hosts/targets and how did you test exactly?

This series is a buildsys refactor; after reviewing I building tested
linux (KVM) and win32 (default, --enable-vhost-user,
--enable-vhost-kernel). I did not test on BSD.
I did not test the running vhost code, simply assumed linking is enough.

> 
>>> ---
>>>  configure             |  2 +-
>>>  hw/net/vhost_net.c    |  4 ++++
>>>  net/Makefile.objs     |  4 +++-
>>>  net/net.c             |  2 +-
>>>  net/vhost-user-stub.c | 23 +++++++++++++++++++++++
>>>  5 files changed, 32 insertions(+), 3 deletions(-)
>>>  create mode 100644 net/vhost-user-stub.c
>>>
>>> diff --git a/configure b/configure
>>> index 0a3c6a7..cda17ef 100755
>>> --- a/configure
>>> +++ b/configure
>>> @@ -6513,7 +6513,7 @@ if test "$vhost_scsi" = "yes" ; then
>>>    echo "CONFIG_VHOST_SCSI=y" >> $config_host_mak
>>>  fi
>>>  if test "$vhost_net" = "yes" -a "$vhost_user" = "yes"; then
>>> -  echo "CONFIG_VHOST_NET_USED=y" >> $config_host_mak
>>> +  echo "CONFIG_VHOST_NET_USER=y" >> $config_host_mak
>>>  fi
>>>  if test "$vhost_crypto" = "yes" ; then
>>>    echo "CONFIG_VHOST_CRYPTO=y" >> $config_host_mak
>>> diff --git a/hw/net/vhost_net.c b/hw/net/vhost_net.c
>>> index b901306..fe6202a 100644
>>> --- a/hw/net/vhost_net.c
>>> +++ b/hw/net/vhost_net.c
>>> @@ -193,6 +193,7 @@ struct vhost_net *vhost_net_init(VhostNetOptions *options)
>>>      }
>>>  
>>>      /* Set sane init value. Override when guest acks. */
>>> +#ifdef CONFIG_VHOST_USER
>>>      if (net->nc->info->type == NET_CLIENT_DRIVER_VHOST_USER) {
>>>          features = vhost_user_get_acked_features(net->nc);
>>>          if (~net->dev.features & features) {
>>> @@ -202,6 +203,7 @@ struct vhost_net *vhost_net_init(VhostNetOptions *options)
>>>              goto fail;
>>>          }
>>>      }
>>> +#endif
>>>  
>>>      vhost_net_ack_features(net, features);
>>>  
>>> @@ -413,10 +415,12 @@ VHostNetState *get_vhost_net(NetClientState *nc)
>>>      case NET_CLIENT_DRIVER_TAP:
>>>          vhost_net = tap_get_vhost_net(nc);
>>>          break;
>>> +#ifdef CONFIG_VHOST_NET_USER
>>>      case NET_CLIENT_DRIVER_VHOST_USER:
>>>          vhost_net = vhost_user_get_vhost_net(nc);
>>>          assert(vhost_net);
>>>          break;
>>> +#endif
>>>      default:
>>>          break;
>>>      }
>>> diff --git a/net/Makefile.objs b/net/Makefile.objs
>>> index b2bf88a..df2b409 100644
>>> --- a/net/Makefile.objs
>>> +++ b/net/Makefile.objs
>>> @@ -3,7 +3,9 @@ common-obj-y += socket.o
>>>  common-obj-y += dump.o
>>>  common-obj-y += eth.o
>>>  common-obj-$(CONFIG_L2TPV3) += l2tpv3.o
>>> -common-obj-$(CONFIG_POSIX) += vhost-user.o
>>> +common-obj-$(call land,$(CONFIG_VIRTIO_NET),$(CONFIG_VHOST_NET_USER)) += vhost-user.o
>>> +common-obj-$(call land,$(call lnot,$(CONFIG_VIRTIO_NET)),$(CONFIG_VHOST_NET_USER)) += vhost-user-stub.o
>>> +common-obj-$(CONFIG_ALL) += vhost-user-stub.o
>>>  common-obj-$(CONFIG_SLIRP) += slirp.o
>>>  common-obj-$(CONFIG_VDE) += vde.o
>>>  common-obj-$(CONFIG_NETMAP) += netmap.o
>>> diff --git a/net/net.c b/net/net.c
>>> index 07c194a..95a74ad 100644
>>> --- a/net/net.c
>>> +++ b/net/net.c
>>> @@ -955,7 +955,7 @@ static int (* const net_client_init_fun[NET_CLIENT_DRIVER__MAX])(
>>>          [NET_CLIENT_DRIVER_BRIDGE]    = net_init_bridge,
>>>  #endif
>>>          [NET_CLIENT_DRIVER_HUBPORT]   = net_init_hubport,
>>> -#ifdef CONFIG_VHOST_NET_USED
>>> +#ifdef CONFIG_VHOST_NET_USER
>>>          [NET_CLIENT_DRIVER_VHOST_USER] = net_init_vhost_user,
>>>  #endif
>>>  #ifdef CONFIG_L2TPV3
>>> diff --git a/net/vhost-user-stub.c b/net/vhost-user-stub.c
>>> new file mode 100644
>>> index 0000000..52ab4e1
>>> --- /dev/null
>>> +++ b/net/vhost-user-stub.c
>>> @@ -0,0 +1,23 @@
>>> +/*
>>> + * vhost-user-stub.c
>>> + *
>>> + * Copyright (c) 2018 Red Hat, Inc.
>>> + *
>>> + * This work is licensed under the terms of the GNU GPL, version 2 or later.
>>> + * See the COPYING file in the top-level directory.
>>> + *
>>> + */
>>> +
>>> +#include "qemu/osdep.h"
>>> +#include "clients.h"
>>> +#include "net/vhost_net.h"
>>> +#include "net/vhost-user.h"
>>> +#include "qemu/error-report.h"
>>> +#include "qapi/error.h"
>>> +
>>> +int net_init_vhost_user(const Netdev *netdev, const char *name,
>>> +                        NetClientState *peer, Error **errp)
>>> +{
>>> +    error_setg(errp, "vhost-user requires frontend driver virtio-net-*");
>>> +    return -1;
>>> +}
>>>

^ permalink raw reply	[flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 0/5] vhost: enable for all targets
  2018-11-26 22:43 ` [Qemu-devel] [PATCH 0/5] vhost: enable for all targets Michael S. Tsirkin
@ 2018-11-27  9:04   ` Paolo Bonzini
  2018-11-27 16:27     ` Michael S. Tsirkin
  0 siblings, 1 reply; 16+ messages in thread
From: Paolo Bonzini @ 2018-11-27  9:04 UTC (permalink / raw)
  To: Michael S. Tsirkin; +Cc: qemu-devel, Marc-André Lureau

On 26/11/18 23:43, Michael S. Tsirkin wrote:
> On Mon, Nov 26, 2018 at 02:20:38PM +0100, Paolo Bonzini wrote:
>> vhost does not have to be supported only if KVM is present, in fact
>> vhost-user does not even need any kind of kernel support.  This series
>> changes this.  The rationale is that, when vhost-user-test.c will be
>> converted to qgraph, it will be able to test vhost-user support
>> for virtio-mmio backend even on x86.
> 
> The reason for limiting it to KVM was very simple:
> it has the same set of problems with ordering
> as mttcg.
> 
> So I guess it's fine but I think we must then limit it
> to when tcg emits fence instructions.
> 
> Otherwise we'll get subtle race conditions.

Got it, I'll take a look.  However, right now you _can_ use vhost with
tcg as long as e.g. qemu-system-x86_64 is compiled with KVM support, so
this patch is not changing anything in this respect; the bug is
independent of this series.

Paolo

^ permalink raw reply	[flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 0/5] vhost: enable for all targets
  2018-11-27  9:04   ` Paolo Bonzini
@ 2018-11-27 16:27     ` Michael S. Tsirkin
  2018-11-27 17:00       ` Paolo Bonzini
  0 siblings, 1 reply; 16+ messages in thread
From: Michael S. Tsirkin @ 2018-11-27 16:27 UTC (permalink / raw)
  To: Paolo Bonzini; +Cc: Marc-André Lureau, qemu-devel

On Tue, Nov 27, 2018 at 10:04:59AM +0100, Paolo Bonzini wrote:
> On 26/11/18 23:43, Michael S. Tsirkin wrote:
> > On Mon, Nov 26, 2018 at 02:20:38PM +0100, Paolo Bonzini wrote:
> >> vhost does not have to be supported only if KVM is present, in fact
> >> vhost-user does not even need any kind of kernel support.  This series
> >> changes this.  The rationale is that, when vhost-user-test.c will be
> >> converted to qgraph, it will be able to test vhost-user support
> >> for virtio-mmio backend even on x86.
> > 
> > The reason for limiting it to KVM was very simple:
> > it has the same set of problems with ordering
> > as mttcg.
> > 
> > So I guess it's fine but I think we must then limit it
> > to when tcg emits fence instructions.
> > 
> > Otherwise we'll get subtle race conditions.
> 
> Got it, I'll take a look.  However, right now you _can_ use vhost with
> tcg as long as e.g. qemu-system-x86_64 is compiled with KVM support, so
> this patch is not changing anything in this respect; the bug is
> independent of this series.
> 
> Paolo
> 

Well qemu built with KVM implies host architecture == guest
architecture, right? That should ensure race conditions do not happen in
a very hacky way.

-- 
MST

^ permalink raw reply	[flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 0/5] vhost: enable for all targets
  2018-11-27 16:27     ` Michael S. Tsirkin
@ 2018-11-27 17:00       ` Paolo Bonzini
  0 siblings, 0 replies; 16+ messages in thread
From: Paolo Bonzini @ 2018-11-27 17:00 UTC (permalink / raw)
  To: Michael S. Tsirkin; +Cc: Marc-André Lureau, qemu-devel

On 27/11/18 17:27, Michael S. Tsirkin wrote:
> On Tue, Nov 27, 2018 at 10:04:59AM +0100, Paolo Bonzini wrote:
>> On 26/11/18 23:43, Michael S. Tsirkin wrote:
>>> On Mon, Nov 26, 2018 at 02:20:38PM +0100, Paolo Bonzini wrote:
>>>> vhost does not have to be supported only if KVM is present, in fact
>>>> vhost-user does not even need any kind of kernel support.  This series
>>>> changes this.  The rationale is that, when vhost-user-test.c will be
>>>> converted to qgraph, it will be able to test vhost-user support
>>>> for virtio-mmio backend even on x86.
>>>
>>> The reason for limiting it to KVM was very simple:
>>> it has the same set of problems with ordering
>>> as mttcg.
>>>
>>> So I guess it's fine but I think we must then limit it
>>> to when tcg emits fence instructions.
>>>
>>> Otherwise we'll get subtle race conditions.
>>
>> Got it, I'll take a look.  However, right now you _can_ use vhost with
>> tcg as long as e.g. qemu-system-x86_64 is compiled with KVM support, so
>> this patch is not changing anything in this respect; the bug is
>> independent of this series.
> 
> Well qemu built with KVM implies host architecture == guest
> architecture, right? That should ensure race conditions do not happen in
> a very hacky way.

It's not enough, event_idx requires smp_mb and single-threaded TCG does
not emit them.  It is enough for smp_rmb/smp_wmb however, that makes sense.

Paolo

^ permalink raw reply	[flat|nested] 16+ messages in thread
end of thread, other threads:[~2018-11-27 17:00 UTC | newest]

Thread overview: 16+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-11-26 13:20 [Qemu-devel] [PATCH 0/5] vhost: enable for all targets Paolo Bonzini
2018-11-26 13:20 ` [Qemu-devel] [PATCH 1/5] vhost-net: move stubs to a separate file Paolo Bonzini
2018-11-26 13:20 ` [Qemu-devel] [PATCH 2/5] vhost-net-user: add stubs for when no virtio-net device is present Paolo Bonzini
2018-11-26 16:45   ` Philippe Mathieu-Daudé
2018-11-26 22:43     ` Michael S. Tsirkin
2018-11-26 22:52       ` Philippe Mathieu-Daudé
2018-11-26 13:20 ` [Qemu-devel] [PATCH 3/5] vhost: restrict Linux dependency to kernel vhost Paolo Bonzini
2018-11-26 16:51   ` Philippe Mathieu-Daudé
2018-11-26 13:20 ` [Qemu-devel] [PATCH 4/5] vhost-net: compile it on all targets that have virtio-net Paolo Bonzini
2018-11-26 16:55   ` Philippe Mathieu-Daudé
2018-11-26 13:20 ` [Qemu-devel] [PATCH 5/5] vhost-net: revamp configure logic Paolo Bonzini
2018-11-26 17:00   ` Philippe Mathieu-Daudé
2018-11-26 22:43 ` [Qemu-devel] [PATCH 0/5] vhost: enable for all targets Michael S. Tsirkin
2018-11-27  9:04   ` Paolo Bonzini
2018-11-27 16:27     ` Michael S. Tsirkin
2018-11-27 17:00       ` Paolo Bonzini

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.