Skip to contentSkip to conversation

vdpa net devices Rx filter change notification with Shadow VQ

PATCH Superseded Eugenio Pérez 1 participant 22 messages Jul 19, 2022 Networking

v5 of 6 Superseded Latest is v6 →

lore
Share & tools
lore
#0 lore raw

v5 cover vdpa net devices Rx filter change notification with Shadow VQ

Control virtqueue is used by networking device for accepting various commands from the driver. It's a must to support advanced configurations.

Rx filtering event is issues by qemu when device's MAC address changed once and the previous one has not been queried by external agents.

Shadow VirtQueue (SVQ) already makes possible tracking the state of virtqueues, effectively intercepting them so qemu can track what regions of memory are dirty because device action and needs migration. However, this does not solve networking device state seen by the driver because CVQ messages, like changes on MAC addresses from the driver.

This series uses SVQ infrastructure to intercept networking control messages used by the device. This way, qemu is able to update VirtIONet device model and react to them. In particular, this series enables rx filter change notification.

This is a prerequisite to achieve net vdpa device with CVQ live migration. It's a stripped down version of [1], with error paths checked and no migration enabled.

First ten patches reorder and clean code base so its easier to apply later ones. No functional change should be noticed from these changes.

Patches from 11 to 15 enable SVQ API to make other parts of qemu to interact with it. In particular, they will be used by vhost-vdpa net to handle CVQ messages.

Patches 16 to 18 enable the update of the virtio-net device model for each CVQ message acknowledged by the device.

Last patches enable x-svq parameter, forbidding device migration since it is not restored in the destination's vdpa device yet. This will be added in later series, using this work.

They could conflict trivially with series [2].

Comments are welcome.
v5:
- Add a volatile access modifier to force compiler to emit the read instruction
- Fix use of undefined variable

v4:
- Add a timeout to vhost_svq_poll to not hold BQL forever

v3:
- Replace SVQElement with SVQDescState

v2:
- (Comments from series [1]).
- Active poll for CVQ answer instead of relay on async used callback
- Do not offer a new buffer to device but reuse qemu's
- Use vhost_svq_add instead of not needed vhost_svq_inject
- Delete used and detach callbacks, not needed anymore
- Embed members of SVQElement in VirtQueueElement
- Reuse the same buffers for all CVQ commands

[1] https://patchwork.kernel.org/project/qemu-devel/cover/[email protected]/
[2] https://lists.nongnu.org/archive/html/qemu-devel/2022-07/msg02984.html

Eugenio Pérez (20):
vhost: move descriptor translation to vhost_svq_vring_write_descs
virtio-net: Expose MAC_TABLE_ENTRIES
virtio-net: Expose ctrl virtqueue logic
vdpa: Avoid compiler to squash reads to used idx
vhost: Reorder vhost_svq_kick
vhost: Move vhost_svq_kick call to vhost_svq_add
vhost: Check for queue full at vhost_svq_add
vhost: Decouple vhost_svq_add from VirtQueueElement
vhost: Add SVQDescState
vhost: Track number of descs in SVQDescState
vhost: add vhost_svq_push_elem
vhost: Expose vhost_svq_add
vhost: add vhost_svq_poll
vhost: Add svq avail_handler callback
vdpa: Export vhost_vdpa_dma_map and unmap calls
vdpa: manual forward CVQ buffers
vdpa: Buffer CVQ support on shadow virtqueue
vdpa: Extract get features part from vhost_vdpa_get_max_queue_pairs
vdpa: Add device migration blocker
vdpa: Add x-svq to NetdevVhostVDPAOptions

 qapi/net.json                      |   9 +-
 hw/virtio/vhost-shadow-virtqueue.h |  52 ++++-
 include/hw/virtio/vhost-vdpa.h     |   8 +
 include/hw/virtio/virtio-net.h     |   7 +
 hw/net/virtio-net.c                |  85 ++++---
 hw/virtio/vhost-shadow-virtqueue.c | 210 ++++++++++++-----
 hw/virtio/vhost-vdpa.c             |  26 ++-
 net/vhost-vdpa.c                   | 357 +++++++++++++++++++++++++++--
 8 files changed, 635 insertions(+), 119 deletions(-)
Signature · 1 line
2.31.1
#1 lore raw

PATCH v5 1/20 vhost: move descriptor translation to vhost_svq_vring_write_descs

It's done for both in and out descriptors so it's better placed here.

---
 hw/virtio/vhost-shadow-virtqueue.c | 38 +++++++++++++++++++++---------
 1 file changed, 27 insertions(+), 11 deletions(-)
hw/virtio/vhost-shadow-virtqueue.c +27 -11
index 56c96ebd13..e2184a4481 100644
@@ -122,17 +122,35 @@ static bool vhost_svq_translate_addr(const VhostShadowVirtqueue *svq,
     return true;
 }
 
-static void vhost_vring_write_descs(VhostShadowVirtqueue *svq, hwaddr *sg,
-                                    const struct iovec *iovec, size_t num,
-                                    bool more_descs, bool write)
+/**
+ * Write descriptors to SVQ vring
+ *
+ * @svq: The shadow virtqueue
+ * @sg: Cache for hwaddr
+ * @iovec: The iovec from the guest
+ * @num: iovec length
+ * @more_descs: True if more descriptors come in the chain
+ * @write: True if they are writeable descriptors
+ *
+ * Return true if success, false otherwise and print error.
+ */
+static bool vhost_svq_vring_write_descs(VhostShadowVirtqueue *svq, hwaddr *sg,
+                                        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;
+    bool ok;
 
     if (num == 0) {
-        return;
+        return true;
+    }
+
+    ok = vhost_svq_translate_addr(svq, sg, iovec, num);
+    if (unlikely(!ok)) {
+        return false;
     }
 
     for (n = 0; n < num; n++) {
@@ -150,6 +168,7 @@ static void vhost_vring_write_descs(VhostShadowVirtqueue *svq, hwaddr *sg,
     }
 
     svq->free_head = le16_to_cpu(svq->desc_next[last]);
+    return true;
 }
 
 static bool vhost_svq_add_split(VhostShadowVirtqueue *svq,
@@ -169,21 +188,18 @@ static bool vhost_svq_add_split(VhostShadowVirtqueue *svq,
         return false;
     }
 
-    ok = vhost_svq_translate_addr(svq, sgs, elem->out_sg, elem->out_num);
+    ok = vhost_svq_vring_write_descs(svq, sgs, elem->out_sg, elem->out_num,
+                                     elem->in_num > 0, false);
     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);
+    ok = vhost_svq_vring_write_descs(svq, sgs, elem->in_sg, elem->in_num, false,
+                                     true);
     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
      * they do sync).
Signature · 1 line
2.31.1
#2 lore raw

PATCH v5 2/20 virtio-net: Expose MAC_TABLE_ENTRIES

In reply to Eugenio Pérez

vhost-vdpa control virtqueue needs to know the maximum entries supported by the virtio-net device, so we know if it is possible to apply the filter.

---
 include/hw/virtio/virtio-net.h | 3 +++
 hw/net/virtio-net.c            | 1 -
 2 files changed, 3 insertions(+), 1 deletion(-)
include/hw/virtio/virtio-net.h +3 -0
index eb87032627..cce1c554f7 100644
@@ -35,6 +35,9 @@ OBJECT_DECLARE_SIMPLE_TYPE(VirtIONet, VIRTIO_NET)
  * and latency. */
 #define TX_BURST 256
 
+/* Maximum VIRTIO_NET_CTRL_MAC_TABLE_SET unicast + multicast entries. */
+#define MAC_TABLE_ENTRIES    64
+
 typedef struct virtio_net_conf
 {
     uint32_t txtimer;
hw/net/virtio-net.c +0 -1
index 7ad948ee7c..f83e96e4ce 100644
@@ -49,7 +49,6 @@
 
 #define VIRTIO_NET_VM_VERSION    11
 
-#define MAC_TABLE_ENTRIES    64
 #define MAX_VLAN    (1 << 12)   /* Per 802.1Q definition */
 
 /* previously fixed value */
Signature · 1 line
2.31.1
#3 lore raw

PATCH v5 3/20 virtio-net: Expose ctrl virtqueue logic

In reply to Eugenio Pérez

This allows external vhost-net devices to modify the state of the VirtIO device model once the vhost-vdpa device has acknowledged the control commands.

---
 include/hw/virtio/virtio-net.h |  4 ++
 hw/net/virtio-net.c            | 84 ++++++++++++++++++++--------------
 2 files changed, 53 insertions(+), 35 deletions(-)
include/hw/virtio/virtio-net.h +4 -0
index cce1c554f7..ef234ffe7e 100644
@@ -221,6 +221,10 @@ struct VirtIONet {
     struct EBPFRSSContext ebpf_rss;
 };
 
+size_t virtio_net_handle_ctrl_iov(VirtIODevice *vdev,
+                                  const struct iovec *in_sg, unsigned in_num,
+                                  const struct iovec *out_sg,
+                                  unsigned out_num);
 void virtio_net_set_netclient_name(VirtIONet *n, const char *name,
                                    const char *type);
 
hw/net/virtio-net.c +49 -35

106 lines not on this page · raw message

Signature · 1 line
2.31.1
#4 lore raw

PATCH v5 4/20 vdpa: Avoid compiler to squash reads to used idx

In reply to Eugenio Pérez

In the next patch we will allow busypolling of this value. The compiler have a running path where shadow_used_idx, last_used_idx, and vring used idx are not modified within the same thread busypolling.

This was not an issue before since we always cleared device event notifier before checking it, and that could act as memory barrier. However, the busypoll needs something similar to kernel READ_ONCE.

Let's add it here, sepparated from the polling.

---
 hw/virtio/vhost-shadow-virtqueue.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
hw/virtio/vhost-shadow-virtqueue.c +2 -1
index e2184a4481..560d07ab36 100644
@@ -327,11 +327,12 @@ static void vhost_handle_guest_kick_notifier(EventNotifier *n)
 
 static bool vhost_svq_more_used(VhostShadowVirtqueue *svq)
 {
+    uint16_t *used_idx = &svq->vring.used->idx;
     if (svq->last_used_idx != svq->shadow_used_idx) {
         return true;
     }
 
-    svq->shadow_used_idx = cpu_to_le16(svq->vring.used->idx);
+    svq->shadow_used_idx = cpu_to_le16(*(volatile uint16_t *)used_idx);
 
     return svq->last_used_idx != svq->shadow_used_idx;
 }
Signature · 1 line
2.31.1
#5 lore raw

PATCH v5 5/20 vhost: Reorder vhost_svq_kick

In reply to Eugenio Pérez

Future code needs to call it from vhost_svq_add.

No functional change intended.

---
 hw/virtio/vhost-shadow-virtqueue.c | 28 ++++++++++++++--------------
 1 file changed, 14 insertions(+), 14 deletions(-)
hw/virtio/vhost-shadow-virtqueue.c +14 -14
index 560d07ab36..043a185b96 100644
@@ -215,6 +215,20 @@ static bool vhost_svq_add_split(VhostShadowVirtqueue *svq,
     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);
+}
+
 /**
  * Add an element to a SVQ.
  *
@@ -235,20 +249,6 @@ static bool vhost_svq_add(VhostShadowVirtqueue *svq, VirtQueueElement *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.
  *
Signature · 1 line
2.31.1
#6 lore raw

PATCH v5 6/20 vhost: Move vhost_svq_kick call to vhost_svq_add

In reply to Eugenio Pérez

The series needs to expose vhost_svq_add with full functionality, including kick

---
 hw/virtio/vhost-shadow-virtqueue.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
hw/virtio/vhost-shadow-virtqueue.c +1 -1
index 043a185b96..e272c3318a 100644
@@ -246,6 +246,7 @@ static bool vhost_svq_add(VhostShadowVirtqueue *svq, VirtQueueElement *elem)
     }
 
     svq->ring_id_maps[qemu_head] = elem;
+    vhost_svq_kick(svq);
     return true;
 }
 
@@ -306,7 +307,6 @@ static void vhost_handle_guest_kick(VhostShadowVirtqueue *svq)
                 /* VQ is broken, just return and ignore any other kicks */
                 return;
             }
-            vhost_svq_kick(svq);
         }
 
         virtio_queue_set_notification(svq->vq, true);
Signature · 1 line
2.31.1
#7 lore raw

PATCH v5 7/20 vhost: Check for queue full at vhost_svq_add

In reply to Eugenio Pérez

The series need to expose vhost_svq_add with full functionality, including checking for full queue.

---
 hw/virtio/vhost-shadow-virtqueue.c | 59 +++++++++++++++++-------------
 1 file changed, 33 insertions(+), 26 deletions(-)
hw/virtio/vhost-shadow-virtqueue.c +33 -26
index e272c3318a..11302ea1f2 100644
@@ -233,21 +233,29 @@ static void vhost_svq_kick(VhostShadowVirtqueue *svq)
  * Add an element to a SVQ.
  *
  * The caller must check that there is enough slots for the new element. It
- * takes ownership of the element: In case of failure, it is free and the SVQ
- * is considered broken.
+ * takes ownership of the element: In case of failure not ENOSPC, it is free.
+ *
+ * Return -EINVAL if element is invalid, -ENOSPC if dev queue is full
  */
-static bool vhost_svq_add(VhostShadowVirtqueue *svq, VirtQueueElement *elem)
+static int vhost_svq_add(VhostShadowVirtqueue *svq, VirtQueueElement *elem)
 {
     unsigned qemu_head;
-    bool ok = vhost_svq_add_split(svq, elem, &qemu_head);
+    unsigned ndescs = elem->in_num + elem->out_num;
+    bool ok;
+
+    if (unlikely(ndescs > vhost_svq_available_slots(svq))) {
+        return -ENOSPC;
+    }
+
+    ok = vhost_svq_add_split(svq, elem, &qemu_head);
     if (unlikely(!ok)) {
         g_free(elem);
-        return false;
+        return -EINVAL;
     }
 
     svq->ring_id_maps[qemu_head] = elem;
     vhost_svq_kick(svq);
-    return true;
+    return 0;
 }
 
 /**
@@ -274,7 +282,7 @@ static void vhost_handle_guest_kick(VhostShadowVirtqueue *svq)
 
         while (true) {
             VirtQueueElement *elem;
-            bool ok;
+            int r;
 
             if (svq->next_guest_avail_elem) {
                 elem = g_steal_pointer(&svq->next_guest_avail_elem);
@@ -286,25 +294,24 @@ static void vhost_handle_guest_kick(VhostShadowVirtqueue *svq)
                 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 */
+            r = vhost_svq_add(svq, elem);
+            if (unlikely(r != 0)) {
+                if (r == -ENOSPC) {
+                    /*
+                     * 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 kicks
+                     * until some elements are used.
+                     */
+                    svq->next_guest_avail_elem = elem;
+                }
+
+                /* VQ is full or broken, just return and ignore kicks */
                 return;
             }
         }
Signature · 1 line
2.31.1
#8 lore raw

PATCH v5 8/20 vhost: Decouple vhost_svq_add from VirtQueueElement

In reply to Eugenio Pérez

VirtQueueElement comes from the guest, but we're heading SVQ to be able to modify the element presented to the device without the guest's knowledge.

To do so, make SVQ accept sg buffers directly, instead of using VirtQueueElement.

Add vhost_svq_add_element to maintain element convenience.

---
 hw/virtio/vhost-shadow-virtqueue.c | 33 ++++++++++++++++++++----------
 1 file changed, 22 insertions(+), 11 deletions(-)
hw/virtio/vhost-shadow-virtqueue.c +22 -11
index 11302ea1f2..e3afa2ba44 100644
@@ -172,30 +172,31 @@ static bool vhost_svq_vring_write_descs(VhostShadowVirtqueue *svq, hwaddr *sg,
 }
 
 static bool vhost_svq_add_split(VhostShadowVirtqueue *svq,
-                                VirtQueueElement *elem, unsigned *head)
+                                const struct iovec *out_sg, size_t out_num,
+                                const struct iovec *in_sg, size_t in_num,
+                                unsigned *head)
 {
     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));
+    g_autofree hwaddr *sgs = g_new(hwaddr, MAX(out_num, in_num));
 
     *head = svq->free_head;
 
     /* We need some descriptors here */
-    if (unlikely(!elem->out_num && !elem->in_num)) {
+    if (unlikely(!out_num && !in_num)) {
         qemu_log_mask(LOG_GUEST_ERROR,
                       "Guest provided element with no descriptors");
         return false;
     }
 
-    ok = vhost_svq_vring_write_descs(svq, sgs, elem->out_sg, elem->out_num,
-                                     elem->in_num > 0, false);
+    ok = vhost_svq_vring_write_descs(svq, sgs, out_sg, out_num, in_num > 0,
+                                     false);
     if (unlikely(!ok)) {
         return false;
     }
 
-    ok = vhost_svq_vring_write_descs(svq, sgs, elem->in_sg, elem->in_num, false,
-                                     true);
+    ok = vhost_svq_vring_write_descs(svq, sgs, in_sg, in_num, false, true);
     if (unlikely(!ok)) {
         return false;
     }
@@ -237,17 +238,19 @@ static void vhost_svq_kick(VhostShadowVirtqueue *svq)
  *
  * Return -EINVAL if element is invalid, -ENOSPC if dev queue is full
  */
-static int vhost_svq_add(VhostShadowVirtqueue *svq, VirtQueueElement *elem)
+static int vhost_svq_add(VhostShadowVirtqueue *svq, const struct iovec *out_sg,
+                          size_t out_num, const struct iovec *in_sg,
+                          size_t in_num, VirtQueueElement *elem)
 {
     unsigned qemu_head;
-    unsigned ndescs = elem->in_num + elem->out_num;
+    unsigned ndescs = in_num + out_num;
     bool ok;
 
     if (unlikely(ndescs > vhost_svq_available_slots(svq))) {
         return -ENOSPC;
     }
 
-    ok = vhost_svq_add_split(svq, elem, &qemu_head);
+    ok = vhost_svq_add_split(svq, out_sg, out_num, in_sg, in_num, &qemu_head);
     if (unlikely(!ok)) {
         g_free(elem);
         return -EINVAL;
@@ -258,6 +261,14 @@ static int vhost_svq_add(VhostShadowVirtqueue *svq, VirtQueueElement *elem)
     return 0;
 }
 
+/* Convenience wrapper to add a guest's element to SVQ */
+static int vhost_svq_add_element(VhostShadowVirtqueue *svq,
+                                 VirtQueueElement *elem)
+{
+    return vhost_svq_add(svq, elem->out_sg, elem->out_num, elem->in_sg,
+                         elem->in_num, elem);
+}
+
 /**
  * Forward available buffers.
  *
@@ -294,7 +305,7 @@ static void vhost_handle_guest_kick(VhostShadowVirtqueue *svq)
                 break;
             }
 
-            r = vhost_svq_add(svq, elem);
+            r = vhost_svq_add_element(svq, elem);
             if (unlikely(r != 0)) {
                 if (r == -ENOSPC) {
                     /*
Signature · 1 line
2.31.1
#9 lore raw

PATCH v5 9/20 vhost: Add SVQDescState

In reply to Eugenio Pérez

This will allow SVQ to add context to the different queue elements.

This patch only store the actual element, no functional change intended.

---
 hw/virtio/vhost-shadow-virtqueue.h |  8 ++++++--
 hw/virtio/vhost-shadow-virtqueue.c | 16 ++++++++--------
 2 files changed, 14 insertions(+), 10 deletions(-)
hw/virtio/vhost-shadow-virtqueue.h +6 -2
index c132c994e9..d646c35054 100644
@@ -15,6 +15,10 @@
 #include "standard-headers/linux/vhost_types.h"
 #include "hw/virtio/vhost-iova-tree.h"
 
+typedef struct SVQDescState {
+    VirtQueueElement *elem;
+} SVQDescState;
+
 /* Shadow virtqueue to relay notifications */
 typedef struct VhostShadowVirtqueue {
     /* Shadow vring */
@@ -47,8 +51,8 @@ typedef struct VhostShadowVirtqueue {
     /* IOVA mapping */
     VhostIOVATree *iova_tree;
 
-    /* Map for use the guest's descriptors */
-    VirtQueueElement **ring_id_maps;
+    /* SVQ vring descriptors state */
+    SVQDescState *desc_state;
 
     /* Next VirtQueue element that guest made available */
     VirtQueueElement *next_guest_avail_elem;
hw/virtio/vhost-shadow-virtqueue.c +8 -8
index e3afa2ba44..e4c09e2d88 100644
@@ -256,7 +256,7 @@ static int vhost_svq_add(VhostShadowVirtqueue *svq, const struct iovec *out_sg,
         return -EINVAL;
     }
 
-    svq->ring_id_maps[qemu_head] = elem;
+    svq->desc_state[qemu_head].elem = elem;
     vhost_svq_kick(svq);
     return 0;
 }
@@ -411,21 +411,21 @@ static VirtQueueElement *vhost_svq_get_buf(VhostShadowVirtqueue *svq,
         return NULL;
     }
 
-    if (unlikely(!svq->ring_id_maps[used_elem.id])) {
+    if (unlikely(!svq->desc_state[used_elem.id].elem)) {
         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;
     }
 
-    num = svq->ring_id_maps[used_elem.id]->in_num +
-          svq->ring_id_maps[used_elem.id]->out_num;
+    num = svq->desc_state[used_elem.id].elem->in_num +
+          svq->desc_state[used_elem.id].elem->out_num;
     last_used_chain = vhost_svq_last_desc_of_chain(svq, num, used_elem.id);
     svq->desc_next[last_used_chain] = svq->free_head;
     svq->free_head = used_elem.id;
 
     *len = used_elem.len;
-    return g_steal_pointer(&svq->ring_id_maps[used_elem.id]);
+    return g_steal_pointer(&svq->desc_state[used_elem.id].elem);
 }
 
 static void vhost_svq_flush(VhostShadowVirtqueue *svq,
@@ -595,7 +595,7 @@ void vhost_svq_start(VhostShadowVirtqueue *svq, VirtIODevice *vdev,
     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);
+    svq->desc_state = g_new0(SVQDescState, svq->vring.num);
     svq->desc_next = g_new0(uint16_t, svq->vring.num);
     for (unsigned i = 0; i < svq->vring.num - 1; i++) {
         svq->desc_next[i] = cpu_to_le16(i + 1);
@@ -620,7 +620,7 @@ void vhost_svq_stop(VhostShadowVirtqueue *svq)
 
     for (unsigned i = 0; i < svq->vring.num; ++i) {
         g_autofree VirtQueueElement *elem = NULL;
-        elem = g_steal_pointer(&svq->ring_id_maps[i]);
+        elem = g_steal_pointer(&svq->desc_state[i].elem);
         if (elem) {
             virtqueue_detach_element(svq->vq, elem, 0);
         }
@@ -632,7 +632,7 @@ void vhost_svq_stop(VhostShadowVirtqueue *svq)
     }
     svq->vq = NULL;
     g_free(svq->desc_next);
-    g_free(svq->ring_id_maps);
+    g_free(svq->desc_state);
     qemu_vfree(svq->vring.desc);
     qemu_vfree(svq->vring.used);
 }
Signature · 1 line
2.31.1
#10 lore raw

PATCH v5 10/20 vhost: Track number of descs in SVQDescState

In reply to Eugenio Pérez

A guest's buffer continuos on GPA may need multiple descriptors on qemu's VA, so SVQ should track its length sepparatedly.

---
 hw/virtio/vhost-shadow-virtqueue.h | 6 ++++++
 hw/virtio/vhost-shadow-virtqueue.c | 4 ++--
 2 files changed, 8 insertions(+), 2 deletions(-)
hw/virtio/vhost-shadow-virtqueue.h +6 -0
index d646c35054..5c7e7cbab6 100644
@@ -17,6 +17,12 @@
 
 typedef struct SVQDescState {
     VirtQueueElement *elem;
+
+    /*
+     * Number of descriptors exposed to the device. May or may not match
+     * guest's
+     */
+    unsigned int ndescs;
 } SVQDescState;
 
 /* Shadow virtqueue to relay notifications */
hw/virtio/vhost-shadow-virtqueue.c +2 -2
index e4c09e2d88..8314405e01 100644
@@ -257,6 +257,7 @@ static int vhost_svq_add(VhostShadowVirtqueue *svq, const struct iovec *out_sg,
     }
 
     svq->desc_state[qemu_head].elem = elem;
+    svq->desc_state[qemu_head].ndescs = ndescs;
     vhost_svq_kick(svq);
     return 0;
 }
@@ -418,8 +419,7 @@ static VirtQueueElement *vhost_svq_get_buf(VhostShadowVirtqueue *svq,
         return NULL;
     }
 
-    num = svq->desc_state[used_elem.id].elem->in_num +
-          svq->desc_state[used_elem.id].elem->out_num;
+    num = svq->desc_state[used_elem.id].ndescs;
     last_used_chain = vhost_svq_last_desc_of_chain(svq, num, used_elem.id);
     svq->desc_next[last_used_chain] = svq->free_head;
     svq->free_head = used_elem.id;
Signature · 1 line
2.31.1
#11 lore raw

PATCH v5 11/20 vhost: add vhost_svq_push_elem

In reply to Eugenio Pérez

This function allows external SVQ users to return guest's available buffers.

---
 hw/virtio/vhost-shadow-virtqueue.h |  3 +++
 hw/virtio/vhost-shadow-virtqueue.c | 16 ++++++++++++++++
 2 files changed, 19 insertions(+)
hw/virtio/vhost-shadow-virtqueue.h +3 -0
index 5c7e7cbab6..d9fc1f1799 100644
@@ -84,6 +84,9 @@ typedef struct VhostShadowVirtqueue {
 
 bool vhost_svq_valid_features(uint64_t features, Error **errp);
 
+void vhost_svq_push_elem(VhostShadowVirtqueue *svq,
+                         const VirtQueueElement *elem, uint32_t len);
+
 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,
hw/virtio/vhost-shadow-virtqueue.c +16 -0
index 8314405e01..1669b1fcd1 100644
@@ -428,6 +428,22 @@ static VirtQueueElement *vhost_svq_get_buf(VhostShadowVirtqueue *svq,
     return g_steal_pointer(&svq->desc_state[used_elem.id].elem);
 }
 
+/**
+ * Push an element to SVQ, returning it to the guest.
+ */
+void vhost_svq_push_elem(VhostShadowVirtqueue *svq,
+                         const VirtQueueElement *elem, uint32_t len)
+{
+    virtqueue_push(svq->vq, elem, len);
+    if (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);
+    }
+}
+
 static void vhost_svq_flush(VhostShadowVirtqueue *svq,
                             bool check_for_avail_queue)
 {
Signature · 1 line
2.31.1
#12 lore raw

PATCH v5 12/20 vhost: Expose vhost_svq_add

In reply to Eugenio Pérez

This allows external parts of SVQ to forward custom buffers to the device.

---
 hw/virtio/vhost-shadow-virtqueue.h | 3 +++
 hw/virtio/vhost-shadow-virtqueue.c | 6 +++---
 2 files changed, 6 insertions(+), 3 deletions(-)
hw/virtio/vhost-shadow-virtqueue.h +3 -0
index d9fc1f1799..dd78f4bec2 100644
@@ -86,6 +86,9 @@ bool vhost_svq_valid_features(uint64_t features, Error **errp);
 
 void vhost_svq_push_elem(VhostShadowVirtqueue *svq,
                          const VirtQueueElement *elem, uint32_t len);
+int vhost_svq_add(VhostShadowVirtqueue *svq, const struct iovec *out_sg,
+                  size_t out_num, const struct iovec *in_sg, size_t in_num,
+                  VirtQueueElement *elem);
 
 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);
hw/virtio/vhost-shadow-virtqueue.c +3 -3
index 1669b1fcd1..c3a75ca89e 100644
@@ -238,9 +238,9 @@ static void vhost_svq_kick(VhostShadowVirtqueue *svq)
  *
  * Return -EINVAL if element is invalid, -ENOSPC if dev queue is full
  */
-static int vhost_svq_add(VhostShadowVirtqueue *svq, const struct iovec *out_sg,
-                          size_t out_num, const struct iovec *in_sg,
-                          size_t in_num, VirtQueueElement *elem)
+int vhost_svq_add(VhostShadowVirtqueue *svq, const struct iovec *out_sg,
+                  size_t out_num, const struct iovec *in_sg, size_t in_num,
+                  VirtQueueElement *elem)
 {
     unsigned qemu_head;
     unsigned ndescs = in_num + out_num;
Signature · 1 line
2.31.1
#13 lore raw

PATCH v5 13/20 vhost: add vhost_svq_poll

In reply to Eugenio Pérez

It allows the Shadow Control VirtQueue to wait for the device to use the available buffers.

---
 hw/virtio/vhost-shadow-virtqueue.h |  1 +
 hw/virtio/vhost-shadow-virtqueue.c | 27 +++++++++++++++++++++++++++
 2 files changed, 28 insertions(+)
hw/virtio/vhost-shadow-virtqueue.h +1 -0
index dd78f4bec2..cf442f7dea 100644
@@ -89,6 +89,7 @@ void vhost_svq_push_elem(VhostShadowVirtqueue *svq,
 int vhost_svq_add(VhostShadowVirtqueue *svq, const struct iovec *out_sg,
                   size_t out_num, const struct iovec *in_sg, size_t in_num,
                   VirtQueueElement *elem);
+size_t vhost_svq_poll(VhostShadowVirtqueue *svq);
 
 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);
hw/virtio/vhost-shadow-virtqueue.c +27 -0
index c3a75ca89e..cc2ee4780d 100644
@@ -485,6 +485,33 @@ static void vhost_svq_flush(VhostShadowVirtqueue *svq,
     } while (!vhost_svq_enable_notification(svq));
 }
 
+/**
+ * Poll the SVQ for one device used buffer.
+ *
+ * This function race with main event loop SVQ polling, so extra
+ * synchronization is needed.
+ *
+ * Return the length written by the device.
+ */
+size_t vhost_svq_poll(VhostShadowVirtqueue *svq)
+{
+    int64_t start_us = g_get_monotonic_time();
+    do {
+        uint32_t len;
+        VirtQueueElement *elem = vhost_svq_get_buf(svq, &len);
+        if (elem) {
+            return len;
+        }
+
+        if (unlikely(g_get_monotonic_time() - start_us > 10e6)) {
+            return 0;
+        }
+
+        /* Make sure we read new used_idx */
+        smp_rmb();
+    } while (true);
+}
+
 /**
  * Forward used buffers.
  *
Signature · 1 line
2.31.1
#14 lore raw

PATCH v5 14/20 vhost: Add svq avail_handler callback

In reply to Eugenio Pérez

This allows external handlers to be aware of new buffers that the guest places in the virtqueue.

When this callback is defined the ownership of the guest's virtqueue element is transferred to the callback. This means that if the user wants to forward the descriptor it needs to manually inject it. The callback is also free to process the command by itself and use the element with svq_push.

---
 hw/virtio/vhost-shadow-virtqueue.h | 31 +++++++++++++++++++++++++++++-
 hw/virtio/vhost-shadow-virtqueue.c | 14 ++++++++++++--
 hw/virtio/vhost-vdpa.c             |  3 ++-
 3 files changed, 44 insertions(+), 4 deletions(-)
hw/virtio/vhost-shadow-virtqueue.h +30 -1
index cf442f7dea..d04c34a589 100644
@@ -25,6 +25,27 @@ typedef struct SVQDescState {
     unsigned int ndescs;
 } SVQDescState;
 
+typedef struct VhostShadowVirtqueue VhostShadowVirtqueue;
+
+/**
+ * Callback to handle an avail buffer.
+ *
+ * @svq:  Shadow virtqueue
+ * @elem:  Element placed in the queue by the guest
+ * @vq_callback_opaque:  Opaque
+ *
+ * Returns 0 if the vq is running as expected.
+ *
+ * Note that ownership of elem is transferred to the callback.
+ */
+typedef int (*VirtQueueAvailCallback)(VhostShadowVirtqueue *svq,
+                                      VirtQueueElement *elem,
+                                      void *vq_callback_opaque);
+
+typedef struct VhostShadowVirtqueueOps {
+    VirtQueueAvailCallback avail_handler;
+} VhostShadowVirtqueueOps;
+
 /* Shadow virtqueue to relay notifications */
 typedef struct VhostShadowVirtqueue {
     /* Shadow vring */
@@ -69,6 +90,12 @@ typedef struct VhostShadowVirtqueue {
      */
     uint16_t *desc_next;
 
+    /* Caller callbacks */
+    const VhostShadowVirtqueueOps *ops;
+
+    /* Caller callbacks opaque */
+    void *ops_opaque;
+
     /* Next head to expose to the device */
     uint16_t shadow_avail_idx;
 
@@ -102,7 +129,9 @@ void vhost_svq_start(VhostShadowVirtqueue *svq, VirtIODevice *vdev,
                      VirtQueue *vq);
 void vhost_svq_stop(VhostShadowVirtqueue *svq);
 
-VhostShadowVirtqueue *vhost_svq_new(VhostIOVATree *iova_tree);
+VhostShadowVirtqueue *vhost_svq_new(VhostIOVATree *iova_tree,
+                                    const VhostShadowVirtqueueOps *ops,
+                                    void *ops_opaque);
 
 void vhost_svq_free(gpointer vq);
 G_DEFINE_AUTOPTR_CLEANUP_FUNC(VhostShadowVirtqueue, vhost_svq_free);
hw/virtio/vhost-shadow-virtqueue.c +12 -2
index cc2ee4780d..e4956728dd 100644
@@ -306,7 +306,11 @@ static void vhost_handle_guest_kick(VhostShadowVirtqueue *svq)
                 break;
             }
 
-            r = vhost_svq_add_element(svq, elem);
+            if (svq->ops) {
+                r = svq->ops->avail_handler(svq, elem, svq->ops_opaque);
+            } else {
+                r = vhost_svq_add_element(svq, elem);
+            }
             if (unlikely(r != 0)) {
                 if (r == -ENOSPC) {
                     /*
@@ -685,12 +689,16 @@ void vhost_svq_stop(VhostShadowVirtqueue *svq)
  * shadow methods and file descriptors.
  *
  * @iova_tree: Tree to perform descriptors translations
+ * @ops: SVQ owner callbacks
+ * @ops_opaque: ops opaque pointer
  *
  * Returns the new virtqueue or NULL.
  *
  * In case of error, reason is reported through error_report.
  */
-VhostShadowVirtqueue *vhost_svq_new(VhostIOVATree *iova_tree)
+VhostShadowVirtqueue *vhost_svq_new(VhostIOVATree *iova_tree,
+                                    const VhostShadowVirtqueueOps *ops,
+                                    void *ops_opaque)
 {
     g_autofree VhostShadowVirtqueue *svq = g_new0(VhostShadowVirtqueue, 1);
     int r;
@@ -712,6 +720,8 @@ VhostShadowVirtqueue *vhost_svq_new(VhostIOVATree *iova_tree)
     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;
+    svq->ops = ops;
+    svq->ops_opaque = ops_opaque;
     return g_steal_pointer(&svq);
 
 err_init_hdev_call:
hw/virtio/vhost-vdpa.c +2 -1
index 66f054a12c..0b13e98471 100644
@@ -418,8 +418,9 @@ 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(v->iova_tree);
+        g_autoptr(VhostShadowVirtqueue) svq;
 
+        svq = vhost_svq_new(v->iova_tree, NULL, NULL);
         if (unlikely(!svq)) {
             error_setg(errp, "Cannot create svq %u", n);
             return -1;
Signature · 1 line
2.31.1
#15 lore raw

PATCH v5 15/20 vdpa: Export vhost_vdpa_dma_map and unmap calls

In reply to Eugenio Pérez

Shadow CVQ will copy buffers on qemu VA, so we avoid TOCTOU attacks from the guest that could set a different state in qemu device model and vdpa device.

To do so, it needs to be able to map these new buffers to the device.

---
 include/hw/virtio/vhost-vdpa.h | 4 ++++
 hw/virtio/vhost-vdpa.c         | 7 +++----
 2 files changed, 7 insertions(+), 4 deletions(-)
include/hw/virtio/vhost-vdpa.h +4 -0
index a29dbb3f53..7214eb47dc 100644
@@ -39,4 +39,8 @@ typedef struct vhost_vdpa {
     VhostVDPAHostNotifier notifier[VIRTIO_QUEUE_MAX];
 } VhostVDPA;
 
+int vhost_vdpa_dma_map(struct vhost_vdpa *v, hwaddr iova, hwaddr size,
+                       void *vaddr, bool readonly);
+int vhost_vdpa_dma_unmap(struct vhost_vdpa *v, hwaddr iova, hwaddr size);
+
 #endif
hw/virtio/vhost-vdpa.c +3 -4
index 0b13e98471..96997210be 100644
@@ -71,8 +71,8 @@ static bool vhost_vdpa_listener_skipped_section(MemoryRegionSection *section,
     return false;
 }
 
-static int vhost_vdpa_dma_map(struct vhost_vdpa *v, hwaddr iova, hwaddr size,
-                              void *vaddr, bool readonly)
+int vhost_vdpa_dma_map(struct vhost_vdpa *v, hwaddr iova, hwaddr size,
+                       void *vaddr, bool readonly)
 {
     struct vhost_msg_v2 msg = {};
     int fd = v->device_fd;
@@ -97,8 +97,7 @@ static int vhost_vdpa_dma_map(struct vhost_vdpa *v, hwaddr iova, hwaddr size,
     return ret;
 }
 
-static int vhost_vdpa_dma_unmap(struct vhost_vdpa *v, hwaddr iova,
-                                hwaddr size)
+int vhost_vdpa_dma_unmap(struct vhost_vdpa *v, hwaddr iova, hwaddr size)
 {
     struct vhost_msg_v2 msg = {};
     int fd = v->device_fd;
Signature · 1 line
2.31.1
#16 lore raw

PATCH v5 16/20 vdpa: manual forward CVQ buffers

In reply to Eugenio Pérez

Do a simple forwarding of CVQ buffers, the same work SVQ could do but through callbacks. No functional change intended.

---
 include/hw/virtio/vhost-vdpa.h |  3 ++
 hw/virtio/vhost-vdpa.c         |  3 +-
 net/vhost-vdpa.c               | 58 ++++++++++++++++++++++++++++++++++
 3 files changed, 63 insertions(+), 1 deletion(-)
include/hw/virtio/vhost-vdpa.h +3 -0
index 7214eb47dc..1111d85643 100644
@@ -15,6 +15,7 @@
 #include <gmodule.h>
 
 #include "hw/virtio/vhost-iova-tree.h"
+#include "hw/virtio/vhost-shadow-virtqueue.h"
 #include "hw/virtio/virtio.h"
 #include "standard-headers/linux/vhost_types.h"
 
@@ -35,6 +36,8 @@ typedef struct vhost_vdpa {
     /* IOVA mapping used by the Shadow Virtqueue */
     VhostIOVATree *iova_tree;
     GPtrArray *shadow_vqs;
+    const VhostShadowVirtqueueOps *shadow_vq_ops;
+    void *shadow_vq_ops_opaque;
     struct vhost_dev *dev;
     VhostVDPAHostNotifier notifier[VIRTIO_QUEUE_MAX];
 } VhostVDPA;
hw/virtio/vhost-vdpa.c +2 -1
index 96997210be..beaaa7049a 100644
@@ -419,7 +419,8 @@ static int vhost_vdpa_init_svq(struct vhost_dev *hdev, struct vhost_vdpa *v,
     for (unsigned n = 0; n < hdev->nvqs; ++n) {
         g_autoptr(VhostShadowVirtqueue) svq;
 
-        svq = vhost_svq_new(v->iova_tree, NULL, NULL);
+        svq = vhost_svq_new(v->iova_tree, v->shadow_vq_ops,
+                            v->shadow_vq_ops_opaque);
         if (unlikely(!svq)) {
             error_setg(errp, "Cannot create svq %u", n);
             return -1;
net/vhost-vdpa.c +58 -0

81 lines not on this page · raw message

Signature · 1 line
2.31.1
#17 lore raw

PATCH v5 17/20 vdpa: Buffer CVQ support on shadow virtqueue

In reply to Eugenio Pérez

Introduce the control virtqueue support for vDPA shadow virtqueue. This is needed for advanced networking features like rx filtering.

Virtio-net control VQ copies the descriptors to qemu's VA, so we avoid TOCTOU with the guest's or device's memory every time there is a device model change. Otherwise, the guest could change the memory content in the time between qemu and the device read it.

To demonstrate command handling, VIRTIO_NET_F_CTRL_MACADDR is implemented. If the virtio-net driver changes MAC the virtio-net device model will be updated with the new one, and a rx filtering change event will be raised.

More cvq commands could be added here straightforwardly but they have not been tested.

---
 net/vhost-vdpa.c | 213 +++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 205 insertions(+), 8 deletions(-)
net/vhost-vdpa.c +205 -8
index 2e3b6b10d8..502f6f9d3e 100644
@@ -33,6 +33,9 @@ typedef struct VhostVDPAState {
     NetClientState nc;
     struct vhost_vdpa vhost_vdpa;
     VHostNetState *vhost_net;
+
+    /* Control commands shadow buffers */
+    void *cvq_cmd_out_buffer, *cvq_cmd_in_buffer;
     bool started;
 } VhostVDPAState;
 
@@ -131,6 +134,8 @@ static void vhost_vdpa_cleanup(NetClientState *nc)
 {
     VhostVDPAState *s = DO_UPCAST(VhostVDPAState, nc, nc);
 
+    qemu_vfree(s->cvq_cmd_out_buffer);
+    qemu_vfree(s->cvq_cmd_in_buffer);
     if (s->vhost_net) {
         vhost_net_cleanup(s->vhost_net);
         g_free(s->vhost_net);
@@ -190,24 +195,191 @@ static NetClientInfo net_vhost_vdpa_info = {
         .check_peer_type = vhost_vdpa_check_peer_type,
 };
 
+static void vhost_vdpa_cvq_unmap_buf(struct vhost_vdpa *v, void *addr)
+{
+    VhostIOVATree *tree = v->iova_tree;
+    DMAMap needle = {
+        /*
+         * No need to specify size or to look for more translations since
+         * this contiguous chunk was allocated by us.
+         */
+        .translated_addr = (hwaddr)(uintptr_t)addr,
+    };
+    const DMAMap *map = vhost_iova_tree_find_iova(tree, &needle);
+    int r;
+
+    if (unlikely(!map)) {
+        error_report("Cannot locate expected map");
+        return;
+    }
+
+    r = vhost_vdpa_dma_unmap(v, map->iova, map->size + 1);
+    if (unlikely(r != 0)) {
+        error_report("Device cannot unmap: %s(%d)", g_strerror(r), r);
+    }
+
+    vhost_iova_tree_remove(tree, map);
+}
+
+static size_t vhost_vdpa_net_cvq_cmd_len(void)
+{
+    /*
+     * MAC_TABLE_SET is the ctrl command that produces the longer out buffer.
+     * In buffer is always 1 byte, so it should fit here
+     */
+    return sizeof(struct virtio_net_ctrl_hdr) +
+           2 * sizeof(struct virtio_net_ctrl_mac) +
+           MAC_TABLE_ENTRIES * ETH_ALEN;
+}
+
+static size_t vhost_vdpa_net_cvq_cmd_page_len(void)
+{
+    return ROUND_UP(vhost_vdpa_net_cvq_cmd_len(), qemu_real_host_page_size());
+}
+
+/** Copy and map a guest buffer. */
+static bool vhost_vdpa_cvq_map_buf(struct vhost_vdpa *v,
+                                   const struct iovec *out_data,
+                                   size_t out_num, size_t data_len, void *buf,
+                                   size_t *written, bool write)
+{
+    DMAMap map = {};
+    int r;
+
+    if (unlikely(!data_len)) {
+        qemu_log_mask(LOG_GUEST_ERROR, "%s: invalid legnth of %s buffer\n",
+                      __func__, write ? "in" : "out");
+        return false;
+    }
+
+    *written = iov_to_buf(out_data, out_num, 0, buf, data_len);
+    map.translated_addr = (hwaddr)(uintptr_t)buf;
+    map.size = vhost_vdpa_net_cvq_cmd_page_len() - 1;
+    map.perm = write ? IOMMU_RW : IOMMU_RO,
+    r = vhost_iova_tree_map_alloc(v->iova_tree, &map);
+    if (unlikely(r != IOVA_OK)) {
+        error_report("Cannot map injected element");
+        return false;
+    }
+
+    r = vhost_vdpa_dma_map(v, map.iova, vhost_vdpa_net_cvq_cmd_page_len(), buf,
+                           !write);
+    if (unlikely(r < 0)) {
+        goto dma_map_err;
+    }
+
+    return true;
+
+dma_map_err:
+    vhost_iova_tree_remove(v->iova_tree, &map);
+    return false;
+}
+
 /**
- * Forward buffer for the moment.
+ * Copy the guest element into a dedicated buffer suitable to be sent to NIC
+ *
+ * @iov: [0] is the out buffer, [1] is the in one
+ */
+static bool vhost_vdpa_net_cvq_map_elem(VhostVDPAState *s,
+                                        VirtQueueElement *elem,
+                                        struct iovec *iov)
+{
+    size_t in_copied;
+    bool ok;
+
+    iov[0].iov_base = s->cvq_cmd_out_buffer;
+    ok = vhost_vdpa_cvq_map_buf(&s->vhost_vdpa, elem->out_sg, elem->out_num,
+                                vhost_vdpa_net_cvq_cmd_len(), iov[0].iov_base,
+                                &iov[0].iov_len, false);
+    if (unlikely(!ok)) {
+        return false;
+    }
+
+    iov[1].iov_base = s->cvq_cmd_in_buffer;
+    ok = vhost_vdpa_cvq_map_buf(&s->vhost_vdpa, NULL, 0,
+                                sizeof(virtio_net_ctrl_ack), iov[1].iov_base,
+                                &in_copied, true);
+    if (unlikely(!ok)) {
+        vhost_vdpa_cvq_unmap_buf(&s->vhost_vdpa, s->cvq_cmd_out_buffer);
+        return false;
+    }
+
+    iov[1].iov_len = sizeof(virtio_net_ctrl_ack);
+    return true;
+}
+
+/**
+ * Do not forward commands not supported by SVQ. Otherwise, the device could
+ * accept it and qemu would not know how to update the device model.
+ */
+static bool vhost_vdpa_net_cvq_validate_cmd(const struct iovec *out,
+                                            size_t out_num)
+{
+    struct virtio_net_ctrl_hdr ctrl;
+    size_t n;
+
+    n = iov_to_buf(out, out_num, 0, &ctrl, sizeof(ctrl));
+    if (unlikely(n < sizeof(ctrl))) {
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "%s: invalid legnth of out buffer %zu\n", __func__, n);
+        return false;
+    }
+
+    switch (ctrl.class) {
+    case VIRTIO_NET_CTRL_MAC:
+        switch (ctrl.cmd) {
+        case VIRTIO_NET_CTRL_MAC_ADDR_SET:
+            return true;
+        default:
+            qemu_log_mask(LOG_GUEST_ERROR, "%s: invalid mac cmd %u\n",
+                          __func__, ctrl.cmd);
+        };
+        break;
+    default:
+        qemu_log_mask(LOG_GUEST_ERROR, "%s: invalid control class %u\n",
+                      __func__, ctrl.class);
+    };
+
+    return false;
+}
+
+/**
+ * Validate and copy control virtqueue commands.
+ *
+ * Following QEMU guidelines, we offer a copy of the buffers to the device to
+ * prevent TOCTOU bugs.
  */
 static int vhost_vdpa_net_handle_ctrl_avail(VhostShadowVirtqueue *svq,
                                             VirtQueueElement *elem,
                                             void *opaque)
 {
-    unsigned int n = elem->out_num + elem->in_num;
-    g_autofree struct iovec *dev_buffers = g_new(struct iovec, n);
+    VhostVDPAState *s = opaque;
     size_t in_len, dev_written;
     virtio_net_ctrl_ack status = VIRTIO_NET_ERR;
-    int r;
+    /* out and in buffers sent to the device */
+    struct iovec dev_buffers[2] = {
+        { .iov_base = s->cvq_cmd_out_buffer },
+        { .iov_base = s->cvq_cmd_in_buffer },
+    };
+    /* in buffer used for device model */
+    const struct iovec in = {
+        .iov_base = &status,
+        .iov_len = sizeof(status),
+    };
+    int r = -EINVAL;
+    bool ok;
+
+    ok = vhost_vdpa_net_cvq_map_elem(s, elem, dev_buffers);
+    if (unlikely(!ok)) {
+        goto out;
+    }
 
-    memcpy(dev_buffers, elem->out_sg, elem->out_num);
-    memcpy(dev_buffers + elem->out_num, elem->in_sg, elem->in_num);
+    ok = vhost_vdpa_net_cvq_validate_cmd(&dev_buffers[0], 1);
+    if (unlikely(!ok)) {
+        goto out;
+    }
 
-    r = vhost_svq_add(svq, &dev_buffers[0], elem->out_num, &dev_buffers[1],
-                      elem->in_num, elem);
+    r = vhost_svq_add(svq, &dev_buffers[0], 1, &dev_buffers[1], 1, elem);
     if (unlikely(r != 0)) {
         if (unlikely(r == -ENOSPC)) {
             qemu_log_mask(LOG_GUEST_ERROR, "%s: No space on device queue\n",
@@ -224,6 +396,18 @@ static int vhost_vdpa_net_handle_ctrl_avail(VhostShadowVirtqueue *svq,
     dev_written = vhost_svq_poll(svq);
     if (unlikely(dev_written < sizeof(status))) {
         error_report("Insufficient written data (%zu)", dev_written);
+        goto out;
+    }
+
+    memcpy(&status, dev_buffers[1].iov_base, sizeof(status));
+    if (status != VIRTIO_NET_OK) {
+        goto out;
+    }
+
+    status = VIRTIO_NET_ERR;
+    virtio_net_handle_ctrl_iov(svq->vdev, &in, 1, dev_buffers, 1);
+    if (status != VIRTIO_NET_OK) {
+        error_report("Bad CVQ processing in model");
     }
 
 out:
@@ -234,6 +418,12 @@ out:
     }
     vhost_svq_push_elem(svq, elem, MIN(in_len, sizeof(status)));
     g_free(elem);
+    if (dev_buffers[0].iov_base) {
+        vhost_vdpa_cvq_unmap_buf(&s->vhost_vdpa, dev_buffers[0].iov_base);
+    }
+    if (dev_buffers[1].iov_base) {
+        vhost_vdpa_cvq_unmap_buf(&s->vhost_vdpa, dev_buffers[1].iov_base);
+    }
     return r;
 }
 
@@ -266,6 +456,13 @@ static NetClientState *net_vhost_vdpa_init(NetClientState *peer,
     s->vhost_vdpa.device_fd = vdpa_device_fd;
     s->vhost_vdpa.index = queue_pair_index;
     if (!is_datapath) {
+        s->cvq_cmd_out_buffer = qemu_memalign(qemu_real_host_page_size(),
+                                            vhost_vdpa_net_cvq_cmd_page_len());
+        memset(s->cvq_cmd_out_buffer, 0, vhost_vdpa_net_cvq_cmd_page_len());
+        s->cvq_cmd_in_buffer = qemu_memalign(qemu_real_host_page_size(),
+                                            vhost_vdpa_net_cvq_cmd_page_len());
+        memset(s->cvq_cmd_in_buffer, 0, vhost_vdpa_net_cvq_cmd_page_len());
+
         s->vhost_vdpa.shadow_vq_ops = &vhost_vdpa_net_svq_ops;
         s->vhost_vdpa.shadow_vq_ops_opaque = s;
     }
Signature · 1 line
2.31.1
#18 lore raw

PATCH v5 18/20 vdpa: Extract get features part from vhost_vdpa_get_max_queue_pairs

In reply to Eugenio Pérez

To know the device features is needed for CVQ SVQ, so SVQ knows if it can handle all commands or not. Extract from vhost_vdpa_get_max_queue_pairs so we can reuse it.

---
 net/vhost-vdpa.c | 30 ++++++++++++++++++++----------
 1 file changed, 20 insertions(+), 10 deletions(-)
net/vhost-vdpa.c +20 -10
index 502f6f9d3e..6e3e9f312a 100644
@@ -474,20 +474,24 @@ static NetClientState *net_vhost_vdpa_init(NetClientState *peer,
     return nc;
 }
 
-static int vhost_vdpa_get_max_queue_pairs(int fd, int *has_cvq, Error **errp)
+static int vhost_vdpa_get_features(int fd, uint64_t *features, Error **errp)
+{
+    int ret = ioctl(fd, VHOST_GET_FEATURES, features);
+    if (unlikely(ret < 0)) {
+        error_setg_errno(errp, errno,
+                         "Fail to query features from vhost-vDPA device");
+    }
+    return ret;
+}
+
+static int vhost_vdpa_get_max_queue_pairs(int fd, uint64_t features,
+                                          int *has_cvq, Error **errp)
 {
     unsigned long config_size = offsetof(struct vhost_vdpa_config, buf);
     g_autofree struct vhost_vdpa_config *config = NULL;
     __virtio16 *max_queue_pairs;
-    uint64_t features;
     int ret;
 
-    ret = ioctl(fd, VHOST_GET_FEATURES, &features);
-    if (ret) {
-        error_setg(errp, "Fail to query features from vhost-vDPA device");
-        return ret;
-    }
-
     if (features & (1 << VIRTIO_NET_F_CTRL_VQ)) {
         *has_cvq = 1;
     } else {
@@ -517,10 +521,11 @@ int net_init_vhost_vdpa(const Netdev *netdev, const char *name,
                         NetClientState *peer, Error **errp)
 {
     const NetdevVhostVDPAOptions *opts;
+    uint64_t features;
     int vdpa_device_fd;
     g_autofree NetClientState **ncs = NULL;
     NetClientState *nc;
-    int queue_pairs, i, has_cvq = 0;
+    int queue_pairs, r, i, has_cvq = 0;
 
     assert(netdev->type == NET_CLIENT_DRIVER_VHOST_VDPA);
     opts = &netdev->u.vhost_vdpa;
@@ -534,7 +539,12 @@ int net_init_vhost_vdpa(const Netdev *netdev, const char *name,
         return -errno;
     }
 
-    queue_pairs = vhost_vdpa_get_max_queue_pairs(vdpa_device_fd,
+    r = vhost_vdpa_get_features(vdpa_device_fd, &features, errp);
+    if (unlikely(r < 0)) {
+        return r;
+    }
+
+    queue_pairs = vhost_vdpa_get_max_queue_pairs(vdpa_device_fd, features,
                                                  &has_cvq, errp);
     if (queue_pairs < 0) {
         qemu_close(vdpa_device_fd);
Signature · 1 line
2.31.1
#19 lore raw
In reply to Eugenio Pérez
Quoted 29 lines · Eugenio Pérez

Control virtqueue is used by networking device for accepting various commands from the driver. It's a must to support advanced configurations.

Rx filtering event is issues by qemu when device's MAC address changed once and the previous one has not been queried by external agents.

Shadow VirtQueue (SVQ) already makes possible tracking the state of virtqueues, effectively intercepting them so qemu can track what regions of memory are dirty because device action and needs migration. However, this does not solve networking device state seen by the driver because CVQ messages, like changes on MAC addresses from the driver.

This series uses SVQ infrastructure to intercept networking control messages used by the device. This way, qemu is able to update VirtIONet device model and react to them. In particular, this series enables rx filter change notification.

This is a prerequisite to achieve net vdpa device with CVQ live migration. It's a stripped down version of [1], with error paths checked and no migration enabled.

First ten patches reorder and clean code base so its easier to apply later ones. No functional change should be noticed from these changes.

Patches from 11 to 15 enable SVQ API to make other parts of qemu to interact with it. In particular, they will be used by vhost-vdpa net to handle CVQ messages.

Quoted 7 lines

Patches 16 to 18 enable the update of the virtio-net device model for each CVQ message acknowledged by the device.

Last patches enable x-svq parameter, forbidding device migration since it is not restored in the destination's vdpa device yet. This will be added in later series, using this work.

As a suggestion, maybe it is better to omit patches 18 to 20 if we don't merge the CVQ start code, same as with SVQ base. Part of the code is meant to be reverted (migration blocker, actual x-svq parameter), we can develop start and ASID code on top of 01-17.

Thanks!

#20 lore raw

PATCH v5 19/20 vdpa: Add device migration blocker

In reply to Eugenio Pérez

Since the vhost-vdpa device is exposing _F_LOG, adding a migration blocker if it uses CVQ.

However, qemu is able to migrate simple devices with no CVQ as long as they use SVQ. To allow it, add a placeholder error to vhost_vdpa, and only add to vhost_dev when used. vhost_dev machinery place the migration blocker if needed.

---
 include/hw/virtio/vhost-vdpa.h |  1 +
 hw/virtio/vhost-vdpa.c         | 15 +++++++++++++++
 2 files changed, 16 insertions(+)
include/hw/virtio/vhost-vdpa.h +1 -0
index 1111d85643..d10a89303e 100644
@@ -35,6 +35,7 @@ typedef struct vhost_vdpa {
     bool shadow_vqs_enabled;
     /* IOVA mapping used by the Shadow Virtqueue */
     VhostIOVATree *iova_tree;
+    Error *migration_blocker;
     GPtrArray *shadow_vqs;
     const VhostShadowVirtqueueOps *shadow_vq_ops;
     void *shadow_vq_ops_opaque;
hw/virtio/vhost-vdpa.c +15 -0
index beaaa7049a..291cd19054 100644
@@ -20,6 +20,7 @@
 #include "hw/virtio/vhost-shadow-virtqueue.h"
 #include "hw/virtio/vhost-vdpa.h"
 #include "exec/address-spaces.h"
+#include "migration/blocker.h"
 #include "qemu/cutils.h"
 #include "qemu/main-loop.h"
 #include "cpu.h"
@@ -1022,6 +1023,13 @@ static bool vhost_vdpa_svqs_start(struct vhost_dev *dev)
         return true;
     }
 
+    if (v->migration_blocker) {
+        int r = migrate_add_blocker(v->migration_blocker, &err);
+        if (unlikely(r < 0)) {
+            return false;
+        }
+    }
+
     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);
@@ -1064,6 +1072,10 @@ err:
         vhost_svq_stop(svq);
     }
 
+    if (v->migration_blocker) {
+        migrate_del_blocker(v->migration_blocker);
+    }
+
     return false;
 }
 
@@ -1083,6 +1095,9 @@ static bool vhost_vdpa_svqs_stop(struct vhost_dev *dev)
         }
     }
 
+    if (v->migration_blocker) {
+        migrate_del_blocker(v->migration_blocker);
+    }
     return true;
 }
 
Signature · 1 line
2.31.1
#21 lore raw

PATCH v5 20/20 vdpa: Add x-svq to NetdevVhostVDPAOptions

In reply to Eugenio Pérez

Finally offering the possibility to enable SVQ from the command line.

---
 qapi/net.json    |  9 +++++-
 net/vhost-vdpa.c | 72 ++++++++++++++++++++++++++++++++++++++++++++++--
 2 files changed, 77 insertions(+), 4 deletions(-)
qapi/net.json +8 -1
index 9af11e9a3b..75ba2cb989 100644
@@ -445,12 +445,19 @@
 # @queues: number of queues to be created for multiqueue vhost-vdpa
 #          (default: 1)
 #
+# @x-svq: Start device with (experimental) shadow virtqueue. (Since 7.1)
+#         (default: false)
+#
+# Features:
+# @unstable: Member @x-svq is experimental.
+#
 # Since: 5.1
 ##
 { 'struct': 'NetdevVhostVDPAOptions',
   'data': {
     '*vhostdev':     'str',
-    '*queues':       'int' } }
+    '*queues':       'int',
+    '*x-svq':        {'type': 'bool', 'features' : [ 'unstable'] } } }
 
 ##
 # @NetdevVmnetHostOptions:
net/vhost-vdpa.c +69 -3

143 lines not on this page · raw message

Signature · 1 line
2.31.1