[PATCH v3 0/3] Decouple iovec management from virtqueue elements

older
[PATCH v2 02/15] serialise: Split...

Laurent Vivier

18 Mar 2026 18 Mar '26

10:19 a.m.

This series prepares the vhost-user path for multi-buffer support, where a single virtqueue element can use more than one iovec entry. Currently, iovec arrays are tightly coupled to virtqueue elements: callers must pre-initialize each element's in_sg/out_sg pointers before calling vu_queue_pop(), and each element is assumed to own exactly one iovec slot. This makes it impossible for a single element to span multiple iovec entries, which is needed for UDP multi-buffer reception. The series decouples iovec storage from elements in three patches: - Patch 1 passes iovec arrays as separate parameters to vu_queue_pop() and vu_queue_map_desc(), so the caller controls where descriptors are mapped rather than reading them from pre-initialized element fields. - Patch 2 passes the actual remaining out_sg capacity to vu_queue_pop() in vu_handle_tx() instead of a fixed per-element constant, enabling dynamic iovec allocation. - Patch 3 moves iovec pool management into vu_collect(), which now accepts the iovec array and tracks consumed entries across elements with a running counter. This removes vu_set_element() and vu_init_elem() entirely. Callers that still assume one iovec per element assert this invariant explicitly until they are updated for multi-buffer. The follow-up udp-iov_vu series builds on this to implement actual multi-buffer support in the UDP vhost-user path. v3: - rebase and add David's R-b - fix coding style (if) - rename in_num to in_total v2: - in patch 3, use iov_used in iov_truncate() rather than elem_cnt as vu_collect() is now providing the number of iovec collected. Laurent Vivier (3): virtio: Pass iovec arrays as separate parameters to vu_queue_pop() vu_handle_tx: Pass actual remaining out_sg capacity to vu_queue_pop() vu_common: Move iovec management into vu_collect() tcp_vu.c | 25 +++++++++------- udp_vu.c | 21 ++++++++------ virtio.c | 29 ++++++++++++++----- virtio.h | 4 ++- vu_common.c | 82 ++++++++++++++++++++++++----------------------------- vu_common.h | 22 ++------------ 6 files changed, 91 insertions(+), 92 deletions(-) -- 2.53.0

Show replies by date

Laurent Vivier

18 Mar 18 Mar

10:19 a.m.

New subject: [PATCH v3 3/3] vu_common: Move iovec management into vu_collect()

Previously, callers had to pre-initialize virtqueue elements with iovec entries using vu_set_element() or vu_init_elem() before calling vu_collect(). This meant each element owned a fixed, pre-assigned iovec slot. Move the iovec array into vu_collect() as explicit parameters (in_sg, max_in_sg, and in_total), letting it pass the remaining iovec capacity directly to vu_queue_pop(). A running current_iov counter tracks consumed entries across elements, so multiple elements share a single iovec pool. The optional in_total output parameter reports how many iovec entries were consumed, allowing callers to track usage across multiple vu_collect() calls. This removes vu_set_element() and vu_init_elem() which are no longer needed, and is a prerequisite for multi-buffer support where a single virtqueue element can use more than one iovec entry. For now, callers assert the current single-iovec-per-element invariant until they are updated to handle multiple iovecs. Signed-off-by: Laurent Vivier --- tcp_vu.c | 25 +++++++++++--------- udp_vu.c | 21 ++++++++++------- vu_common.c | 68 ++++++++++++++++++++++++----------------------------- vu_common.h | 22 +++-------------- 4 files changed, 60 insertions(+), 76 deletions(-) diff --git a/tcp_vu.c b/tcp_vu.c index fd734e857b3b..5de9eec9b962 100644 --- a/tcp_vu.c +++ b/tcp_vu.c @@ -87,13 +87,13 @@ int tcp_vu_send_flag(const struct ctx *c, struct tcp_tap_conn *conn, int flags) hdrlen = tcp_vu_hdrlen(CONN_V6(conn)); - vu_set_element(&flags_elem[0], NULL, &flags_iov[0]); - elem_cnt = vu_collect(vdev, vq, &flags_elem[0], 1, + &flags_iov[0], 1, NULL, MAX(hdrlen + sizeof(*opts), ETH_ZLEN + VNET_HLEN), NULL); if (elem_cnt != 1) return -1; + ASSERT(flags_elem[0].in_num == 1); ASSERT(flags_elem[0].in_sg[0].iov_len >= MAX(hdrlen + sizeof(*opts), ETH_ZLEN + VNET_HLEN)); @@ -148,9 +148,8 @@ int tcp_vu_send_flag(const struct ctx *c, struct tcp_tap_conn *conn, int flags) nb_ack = 1; if (flags & DUP_ACK) { - vu_set_element(&flags_elem[1], NULL, &flags_iov[1]); - elem_cnt = vu_collect(vdev, vq, &flags_elem[1], 1, + &flags_iov[1], 1, NULL, flags_elem[0].in_sg[0].iov_len, NULL); if (elem_cnt == 1 && flags_elem[1].in_sg[0].iov_len >= @@ -191,8 +190,8 @@ static ssize_t tcp_vu_sock_recv(const struct ctx *c, struct vu_virtq *vq, const struct vu_dev *vdev = c->vdev; struct msghdr mh_sock = { 0 }; uint16_t mss = MSS_GET(conn); + size_t hdrlen, iov_used; int s = conn->sock; - size_t hdrlen; int elem_cnt; ssize_t ret; int i; @@ -201,22 +200,26 @@ static ssize_t tcp_vu_sock_recv(const struct ctx *c, struct vu_virtq *vq, hdrlen = tcp_vu_hdrlen(v6); - vu_init_elem(elem, &iov_vu[DISCARD_IOV_NUM], VIRTQUEUE_MAX_SIZE); - + iov_used = 0; elem_cnt = 0; *head_cnt = 0; - while (fillsize > 0 && elem_cnt < VIRTQUEUE_MAX_SIZE) { + while (fillsize > 0 && elem_cnt < ARRAY_SIZE(elem) && + iov_used < VIRTQUEUE_MAX_SIZE) { + size_t frame_size, dlen, in_total; struct iovec *iov; - size_t frame_size, dlen; int cnt; cnt = vu_collect(vdev, vq, &elem[elem_cnt], - VIRTQUEUE_MAX_SIZE - elem_cnt, + ARRAY_SIZE(elem) - elem_cnt, + &iov_vu[DISCARD_IOV_NUM + iov_used], + VIRTQUEUE_MAX_SIZE - iov_used, &in_total, MAX(MIN(mss, fillsize) + hdrlen, ETH_ZLEN + VNET_HLEN), &frame_size); if (cnt == 0) break; + ASSERT((size_t)cnt == in_total); /* one iovec per element */ + iov_used += in_total; dlen = frame_size - hdrlen; /* reserve space for headers in iov */ @@ -247,7 +250,7 @@ static ssize_t tcp_vu_sock_recv(const struct ctx *c, struct vu_virtq *vq, ret -= already_sent; /* adjust iov number and length of the last iov */ - i = iov_truncate(&iov_vu[DISCARD_IOV_NUM], elem_cnt, ret); + i = iov_truncate(&iov_vu[DISCARD_IOV_NUM], iov_used, ret); /* adjust head count */ while (*head_cnt > 0 && head[*head_cnt - 1] >= i) diff --git a/udp_vu.c b/udp_vu.c index 5effca777e0a..acc18d375a0f 100644 --- a/udp_vu.c +++ b/udp_vu.c @@ -71,9 +71,10 @@ static int udp_vu_sock_recv(const struct ctx *c, struct vu_virtq *vq, int s, bool v6, ssize_t *dlen) { const struct vu_dev *vdev = c->vdev; + int elem_cnt, elem_used, iov_used; struct msghdr msg = { 0 }; - int iov_cnt, iov_used; size_t hdrlen, l2len; + size_t iov_cnt; ASSERT(!c->no_udp); @@ -89,13 +90,14 @@ static int udp_vu_sock_recv(const struct ctx *c, struct vu_virtq *vq, int s, /* compute L2 header length */ hdrlen = udp_vu_hdrlen(v6); - vu_init_elem(elem, iov_vu, VIRTQUEUE_MAX_SIZE); - - iov_cnt = vu_collect(vdev, vq, elem, VIRTQUEUE_MAX_SIZE, - IP_MAX_MTU + ETH_HLEN + VNET_HLEN, NULL); - if (iov_cnt == 0) + elem_cnt = vu_collect(vdev, vq, elem, ARRAY_SIZE(elem), + iov_vu, ARRAY_SIZE(iov_vu), &iov_cnt, + IP_MAX_MTU + ETH_HLEN + VNET_HLEN, NULL); + if (elem_cnt == 0) return -1; + ASSERT((size_t)elem_cnt == iov_cnt); /* one iovec per element */ + /* reserve space for the headers */ ASSERT(iov_vu[0].iov_len >= MAX(hdrlen, ETH_ZLEN + VNET_HLEN)); iov_vu[0].iov_base = (char *)iov_vu[0].iov_base + hdrlen; @@ -107,7 +109,7 @@ static int udp_vu_sock_recv(const struct ctx *c, struct vu_virtq *vq, int s, *dlen = recvmsg(s, &msg, 0); if (*dlen < 0) { - vu_queue_rewind(vq, iov_cnt); + vu_queue_rewind(vq, elem_cnt); return -1; } @@ -116,15 +118,16 @@ static int udp_vu_sock_recv(const struct ctx *c, struct vu_virtq *vq, int s, iov_vu[0].iov_len += hdrlen; iov_used = iov_truncate(iov_vu, iov_cnt, *dlen + hdrlen); + elem_used = iov_used; /* one iovec per element */ /* pad frame to 60 bytes: first buffer is at least ETH_ZLEN long */ l2len = *dlen + hdrlen - VNET_HLEN; vu_pad(&iov_vu[0], l2len); - vu_set_vnethdr(iov_vu[0].iov_base, iov_used); + vu_set_vnethdr(iov_vu[0].iov_base, elem_used); /* release unused buffers */ - vu_queue_rewind(vq, iov_cnt - iov_used); + vu_queue_rewind(vq, elem_cnt - elem_used); return iov_used; } diff --git a/vu_common.c b/vu_common.c index 834a54fe1858..b6e1fa91eec5 100644 --- a/vu_common.c +++ b/vu_common.c @@ -51,28 +51,15 @@ int vu_packet_check_range(struct vdev_memory *memory, return -1; } -/** - * vu_init_elem() - initialize an array of virtqueue elements with 1 iov in each - * @elem: Array of virtqueue elements to initialize - * @iov: Array of iovec to assign to virtqueue element - * @elem_cnt: Number of virtqueue element - */ -void vu_init_elem(struct vu_virtq_element *elem, struct iovec *iov, int elem_cnt) -{ - int i; - - for (i = 0; i < elem_cnt; i++) - vu_set_element(&elem[i], NULL, &iov[i]); -} - /** * vu_collect() - collect virtio buffers from a given virtqueue * @vdev: vhost-user device * @vq: virtqueue to collect from - * @elem: Array of virtqueue element - * each element must be initialized with one iovec entry - * in the in_sg array. + * @elem: Array of @max_elem virtqueue elements * @max_elem: Number of virtqueue elements in the array + * @in_sg: Incoming iovec array for device-writable descriptors + * @max_in_sg: Maximum number of entries in @in_sg + * @in_total: Number of collected entries from @in_sg (output) * @size: Maximum size of the data in the frame * @collected: Collected buffer length, up to @size, set on return * @@ -80,20 +67,21 @@ void vu_init_elem(struct vu_virtq_element *elem, struct iovec *iov, int elem_cnt */ int vu_collect(const struct vu_dev *vdev, struct vu_virtq *vq, struct vu_virtq_element *elem, int max_elem, + struct iovec *in_sg, size_t max_in_sg, size_t *in_total, size_t size, size_t *collected) { size_t current_size = 0; + size_t current_iov = 0; int elem_cnt = 0; - while (current_size < size && elem_cnt < max_elem) { - struct iovec *iov; + while (current_size < size && elem_cnt < max_elem && + current_iov < max_in_sg) { int ret; ret = vu_queue_pop(vdev, vq, &elem[elem_cnt], - elem[elem_cnt].in_sg, - elem[elem_cnt].in_num, - elem[elem_cnt].out_sg, - elem[elem_cnt].out_num); + &in_sg[current_iov], + max_in_sg - current_iov, + NULL, 0); if (ret < 0) break; @@ -103,18 +91,22 @@ int vu_collect(const struct vu_dev *vdev, struct vu_virtq *vq, break; } - iov = &elem[elem_cnt].in_sg[0]; - - if (iov->iov_len > size - current_size) - iov->iov_len = size - current_size; + elem[elem_cnt].in_num = iov_truncate(elem[elem_cnt].in_sg, + elem[elem_cnt].in_num, + size - current_size); - current_size += iov->iov_len; + current_size += iov_size(elem[elem_cnt].in_sg, + elem[elem_cnt].in_num); + current_iov += elem[elem_cnt].in_num; elem_cnt++; if (!vu_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF)) break; } + if (in_total) + *in_total = current_iov; + if (collected) *collected = current_size; @@ -147,8 +139,11 @@ void vu_flush(const struct vu_dev *vdev, struct vu_virtq *vq, { int i; - for (i = 0; i < elem_cnt; i++) - vu_queue_fill(vdev, vq, &elem[i], elem[i].in_sg[0].iov_len, i); + for (i = 0; i < elem_cnt; i++) { + size_t elem_size = iov_size(elem[i].in_sg, elem[i].in_num); + + vu_queue_fill(vdev, vq, &elem[i], elem_size, i); + } vu_queue_flush(vdev, vq, elem_cnt); vu_queue_notify(vdev, vq); @@ -245,7 +240,7 @@ int vu_send_single(const struct ctx *c, const void *buf, size_t size) struct vu_virtq *vq = &vdev->vq[VHOST_USER_RX_QUEUE]; struct vu_virtq_element elem[VIRTQUEUE_MAX_SIZE]; struct iovec in_sg[VIRTQUEUE_MAX_SIZE]; - size_t total; + size_t total, in_total; int elem_cnt; int i; @@ -256,11 +251,10 @@ int vu_send_single(const struct ctx *c, const void *buf, size_t size) return -1; } - vu_init_elem(elem, in_sg, VIRTQUEUE_MAX_SIZE); - size += VNET_HLEN; - elem_cnt = vu_collect(vdev, vq, elem, VIRTQUEUE_MAX_SIZE, size, &total); - if (total < size) { + elem_cnt = vu_collect(vdev, vq, elem, ARRAY_SIZE(elem), in_sg, + ARRAY_SIZE(in_sg), &in_total, size, &total); + if (elem_cnt == 0 || total < size) { debug("vu_send_single: no space to send the data " "elem_cnt %d size %zd", elem_cnt, total); goto err; @@ -271,10 +265,10 @@ int vu_send_single(const struct ctx *c, const void *buf, size_t size) total -= VNET_HLEN; /* copy data from the buffer to the iovec */ - iov_from_buf(in_sg, elem_cnt, VNET_HLEN, buf, total); + iov_from_buf(in_sg, in_total, VNET_HLEN, buf, total); if (*c->pcap) - pcap_iov(in_sg, elem_cnt, VNET_HLEN); + pcap_iov(in_sg, in_total, VNET_HLEN); vu_flush(vdev, vq, elem, elem_cnt); diff --git a/vu_common.h b/vu_common.h index 865d9771fa89..7b060eb6184f 100644 --- a/vu_common.h +++ b/vu_common.h @@ -35,26 +35,10 @@ static inline void *vu_payloadv6(void *base) return (struct ipv6hdr *)vu_ip(base) + 1; } -/** - * vu_set_element() - Initialize a vu_virtq_element - * @elem: Element to initialize - * @out_sg: One out iovec entry to set in elem - * @in_sg: One in iovec entry to set in elem - */ -static inline void vu_set_element(struct vu_virtq_element *elem, - struct iovec *out_sg, struct iovec *in_sg) -{ - elem->out_num = !!out_sg; - elem->out_sg = out_sg; - elem->in_num = !!in_sg; - elem->in_sg = in_sg; -} - -void vu_init_elem(struct vu_virtq_element *elem, struct iovec *iov, - int elem_cnt); int vu_collect(const struct vu_dev *vdev, struct vu_virtq *vq, - struct vu_virtq_element *elem, int max_elem, size_t size, - size_t *collected); + struct vu_virtq_element *elem, int max_elem, + struct iovec *in_sg, size_t max_in_sg, size_t *in_total, + size_t size, size_t *collected); void vu_set_vnethdr(struct virtio_net_hdr_mrg_rxbuf *vnethdr, int num_buffers); void vu_flush(const struct vu_dev *vdev, struct vu_virtq *vq, struct vu_virtq_element *elem, int elem_cnt); -- 2.53.0

Laurent Vivier

10:19 a.m.

New subject: [PATCH v3 2/3] vu_handle_tx: Pass actual remaining out_sg capacity to vu_queue_pop()

In vu_handle_tx(), pass the actual remaining iovec capacity (ARRAY_SIZE(out_sg) - out_sg_count) to vu_queue_pop() rather than a fixed VU_MAX_TX_BUFFER_NB. This enables dynamic allocation of iovec entries to each element rather than reserving a fixed number of slots per descriptor. Signed-off-by: Laurent Vivier Reviewed-by: David Gibson --- vu_common.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) diff --git a/vu_common.c b/vu_common.c index 4d809ac38a4b..834a54fe1858 100644 --- a/vu_common.c +++ b/vu_common.c @@ -20,8 +20,6 @@ #include "migrate.h" #include "epoll_ctl.h" -#define VU_MAX_TX_BUFFER_NB 2 - /** * vu_packet_check_range() - Check if a given memory zone is contained in * a mapped guest memory region @@ -177,13 +175,13 @@ static void vu_handle_tx(struct vu_dev *vdev, int index, count = 0; out_sg_count = 0; - while (count < VIRTQUEUE_MAX_SIZE && - out_sg_count + VU_MAX_TX_BUFFER_NB <= VIRTQUEUE_MAX_SIZE) { - int ret; + while (count < ARRAY_SIZE(elem) && out_sg_count < ARRAY_SIZE(out_sg)) { struct iov_tail data; + int ret; ret = vu_queue_pop(vdev, vq, &elem[count], NULL, 0, - &out_sg[out_sg_count], VU_MAX_TX_BUFFER_NB); + &out_sg[out_sg_count], + ARRAY_SIZE(out_sg) - out_sg_count); if (ret < 0) break; out_sg_count += elem[count].out_num; -- 2.53.0

Laurent Vivier

10:19 a.m.

New subject: [PATCH v3 1/3] virtio: Pass iovec arrays as separate parameters to vu_queue_pop()

Currently vu_queue_pop() and vu_queue_map_desc() read the iovec arrays (in_sg/out_sg) and their sizes (in_num/out_num) from the vu_virtq_element struct. This couples the iovec storage to the element, requiring callers like vu_handle_tx() to pre-initialize the element fields before calling vu_queue_pop(). Pass the iovec arrays and their maximum sizes as separate parameters instead. vu_queue_map_desc() now writes the actual descriptor count and iovec pointers back into the element after mapping, rather than using the element as both input and output. This decouples the iovec storage from the element, which is a prerequisite for multi-buffer support where a single frame can span multiple virtqueue elements sharing a common iovec pool. No functional change. Signed-off-by: Laurent Vivier Reviewed-by: David Gibson --- virtio.c | 29 ++++++++++++++++++++++------- virtio.h | 4 +++- vu_common.c | 14 +++++++------- 3 files changed, 32 insertions(+), 15 deletions(-) diff --git a/virtio.c b/virtio.c index 447137ee83dd..a671163c27a0 100644 --- a/virtio.c +++ b/virtio.c @@ -428,12 +428,18 @@ static bool virtqueue_map_desc(const struct vu_dev *dev, * @vq: Virtqueue * @idx: First descriptor ring entry to map * @elem: Virtqueue element to store descriptor ring iov + * @in_sg: Incoming iovec array for device-writable descriptors + * @max_in_sg: Maximum number of entries in @in_sg + * @out_sg: Outgoing iovec array for device-readable descriptors + * @max_out_sg: Maximum number of entries in @out_sg * * Return: -1 if there is an error, 0 otherwise */ static int vu_queue_map_desc(const struct vu_dev *dev, struct vu_virtq *vq, unsigned int idx, - struct vu_virtq_element *elem) + struct vu_virtq_element *elem, + struct iovec *in_sg, size_t max_in_sg, + struct iovec *out_sg, size_t max_out_sg) { const struct vring_desc *desc = vq->vring.desc; struct vring_desc desc_buf[VIRTQUEUE_MAX_SIZE]; @@ -470,16 +476,16 @@ static int vu_queue_map_desc(const struct vu_dev *dev, /* Collect all the descriptors */ do { if (le16toh(desc[i].flags) & VRING_DESC_F_WRITE) { - if (!virtqueue_map_desc(dev, &in_num, elem->in_sg, - elem->in_num, + if (!virtqueue_map_desc(dev, &in_num, in_sg, + max_in_sg, le64toh(desc[i].addr), le32toh(desc[i].len))) return -1; } else { if (in_num) die("Incorrect order for descriptors"); - if (!virtqueue_map_desc(dev, &out_num, elem->out_sg, - elem->out_num, + if (!virtqueue_map_desc(dev, &out_num, out_sg, + max_out_sg, le64toh(desc[i].addr), le32toh(desc[i].len))) { return -1; @@ -496,7 +502,9 @@ static int vu_queue_map_desc(const struct vu_dev *dev, die("vhost-user: Failed to read descriptor list"); elem->index = idx; + elem->in_sg = in_sg; elem->in_num = in_num; + elem->out_sg = out_sg; elem->out_num = out_num; return 0; @@ -507,11 +515,17 @@ static int vu_queue_map_desc(const struct vu_dev *dev, * @dev: Vhost-user device * @vq: Virtqueue * @elem: Virtqueue element to fill with the entry information + * @in_sg: Incoming iovec array for device-writable descriptors + * @max_in_sg: Maximum number of entries in @in_sg + * @out_sg: Outgoing iovec array for device-readable descriptors + * @max_out_sg: Maximum number of entries in @out_sg * * Return: -1 if there is an error, 0 otherwise */ int vu_queue_pop(const struct vu_dev *dev, struct vu_virtq *vq, - struct vu_virtq_element *elem) + struct vu_virtq_element *elem, + struct iovec *in_sg, size_t max_in_sg, + struct iovec *out_sg, size_t max_out_sg) { unsigned int head; int ret; @@ -535,7 +549,8 @@ int vu_queue_pop(const struct vu_dev *dev, struct vu_virtq *vq, if (vu_has_feature(dev, VIRTIO_RING_F_EVENT_IDX)) vring_set_avail_event(vq, vq->last_avail_idx); - ret = vu_queue_map_desc(dev, vq, head, elem); + ret = vu_queue_map_desc(dev, vq, head, elem, in_sg, max_in_sg, + out_sg, max_out_sg); if (ret < 0) return ret; diff --git a/virtio.h b/virtio.h index d04bbe84e5c4..c7e447d59860 100644 --- a/virtio.h +++ b/virtio.h @@ -188,7 +188,9 @@ static inline bool vu_has_protocol_feature(const struct vu_dev *vdev, void vu_queue_notify(const struct vu_dev *dev, struct vu_virtq *vq); int vu_queue_pop(const struct vu_dev *dev, struct vu_virtq *vq, - struct vu_virtq_element *elem); + struct vu_virtq_element *elem, + struct iovec *in_sg, size_t max_in_sg, + struct iovec *out_sg, size_t max_out_sg); void vu_queue_detach_element(struct vu_virtq *vq); void vu_queue_unpop(struct vu_virtq *vq); bool vu_queue_rewind(struct vu_virtq *vq, unsigned int num); diff --git a/vu_common.c b/vu_common.c index 5f2ce18e5b71..4d809ac38a4b 100644 --- a/vu_common.c +++ b/vu_common.c @@ -91,7 +91,11 @@ int vu_collect(const struct vu_dev *vdev, struct vu_virtq *vq, struct iovec *iov; int ret; - ret = vu_queue_pop(vdev, vq, &elem[elem_cnt]); + ret = vu_queue_pop(vdev, vq, &elem[elem_cnt], + elem[elem_cnt].in_sg, + elem[elem_cnt].in_num, + elem[elem_cnt].out_sg, + elem[elem_cnt].out_num); if (ret < 0) break; @@ -178,12 +182,8 @@ static void vu_handle_tx(struct vu_dev *vdev, int index, int ret; struct iov_tail data; - elem[count].out_num = VU_MAX_TX_BUFFER_NB; - elem[count].out_sg = &out_sg[out_sg_count]; - elem[count].in_num = 0; - elem[count].in_sg = NULL; - - ret = vu_queue_pop(vdev, vq, &elem[count]); + ret = vu_queue_pop(vdev, vq, &elem[count], NULL, 0, + &out_sg[out_sg_count], VU_MAX_TX_BUFFER_NB); if (ret < 0) break; out_sg_count += elem[count].out_num; -- 2.53.0

Stefano Brivio

20 Mar 20 Mar

9:58 p.m.

On Wed, 18 Mar 2026 10:19:38 +0100 Laurent Vivier wrote:

...

This series prepares the vhost-user path for multi-buffer support, where a single virtqueue element can use more than one iovec entry.

Currently, iovec arrays are tightly coupled to virtqueue elements: callers must pre-initialize each element's in_sg/out_sg pointers before calling vu_queue_pop(), and each element is assumed to own exactly one iovec slot. This makes it impossible for a single element to span multiple iovec entries, which is needed for UDP multi-buffer reception.

The series decouples iovec storage from elements in three patches:

- Patch 1 passes iovec arrays as separate parameters to vu_queue_pop() and vu_queue_map_desc(), so the caller controls where descriptors are mapped rather than reading them from pre-initialized element fields.

- Patch 2 passes the actual remaining out_sg capacity to vu_queue_pop() in vu_handle_tx() instead of a fixed per-element constant, enabling dynamic iovec allocation.

- Patch 3 moves iovec pool management into vu_collect(), which now accepts the iovec array and tracks consumed entries across elements with a running counter. This removes vu_set_element() and vu_init_elem() entirely. Callers that still assume one iovec per element assert this invariant explicitly until they are updated for multi-buffer.

The follow-up udp-iov_vu series builds on this to implement actual multi-buffer support in the UDP vhost-user path.

v3: - rebase and add David's R-b - fix coding style (if) - rename in_num to in_total

Applied. I took the liberty to add David's Reviewed-by: back on 3/3 as the only change in 3/3 v3 compared to v2 was actually something he suggested. -- Stefano

David Gibson

21 Mar 21 Mar

8:40 a.m.

On Fri, Mar 20, 2026 at 09:58:17PM +0100, Stefano Brivio wrote:

...

On Wed, 18 Mar 2026 10:19:38 +0100 Laurent Vivier wrote:

...
This series prepares the vhost-user path for multi-buffer support, where a single virtqueue element can use more than one iovec entry.

Currently, iovec arrays are tightly coupled to virtqueue elements: callers must pre-initialize each element's in_sg/out_sg pointers before calling vu_queue_pop(), and each element is assumed to own exactly one iovec slot. This makes it impossible for a single element to span multiple iovec entries, which is needed for UDP multi-buffer reception.

The series decouples iovec storage from elements in three patches:

- Patch 1 passes iovec arrays as separate parameters to vu_queue_pop() and vu_queue_map_desc(), so the caller controls where descriptors are mapped rather than reading them from pre-initialized element fields.

- Patch 2 passes the actual remaining out_sg capacity to vu_queue_pop() in vu_handle_tx() instead of a fixed per-element constant, enabling dynamic iovec allocation.

- Patch 3 moves iovec pool management into vu_collect(), which now accepts the iovec array and tracks consumed entries across elements with a running counter. This removes vu_set_element() and vu_init_elem() entirely. Callers that still assume one iovec per element assert this invariant explicitly until they are updated for multi-buffer.

The follow-up udp-iov_vu series builds on this to implement actual multi-buffer support in the UDP vhost-user path.

v3: - rebase and add David's R-b - fix coding style (if) - rename in_num to in_total

Applied. I took the liberty to add David's Reviewed-by: back on 3/3 as the only change in 3/3 v3 compared to v2 was actually something he suggested.

Oh, sorry, I thought I'd already R-b'ed all of v3. -- David Gibson (he or they) | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you, not the other way | around. http://www.ozlabs.org/~dgibson

Age (days ago)

Last active (days ago)

List overview

Download

5 comments

3 participants

participants (3)

David Gibson
Laurent Vivier
Stefano Brivio