* [net-next rfc v7 0/3] Multiqueue virtio-net
@ 2012-11-27 10:15 Jason Wang
2012-11-27 10:15 ` [net-next rfc v7 1/3] virtio-net: separate fields of sending/receiving queue from virtnet_info Jason Wang
` (2 more replies)
0 siblings, 3 replies; 26+ messages in thread
From: Jason Wang @ 2012-11-27 10:15 UTC (permalink / raw)
To: rusty, mst, krkumar2, virtualization, netdev, linux-kernel
Cc: kvm, bhutchings, jwhan, shiyer, Jason Wang
Hi all:
This series is an update version of multiqueue virtio-net driver based on
Krishna Kumar's work to let virtio-net use multiple rx/tx queues to do the
packets reception and transmission. Please review and comments.
A protype implementation of qemu-kvm support could by found in
git://github.com/jasowang/qemu-kvm-mq.git. To start a guest with two queues, you
could specify the queues parameters to both tap and virtio-net like:
./qemu-kvm -netdev tap,queues=2,... -device virtio-net-pci,queues=2,...
then enable the multiqueue through ethtool by:
ethtool -L eth0 combined 2
Changes from V6:
- Align the implementation with the RFC spec update v5
- Addressing Rusty's comments:
* split the patches
* rename to max_queue_pairs and curr_queue_pairs
* remove the useless status
* fix the hibernation bug
- Addressing Ben's comments:
* check other parameters in ethtool_set_queues
Changes from v5:
- Align the implementation with the RFC spec update v4
- Switch the mode between single mode and multiqueue mode without reset
- Remove the 256 limitation of queues
- Use helpers to do the mapping between virtqueues and tx/rx queues
- Use commbined channels instead of separated rx/tx queus when do the queue
number configuartion
- Other coding style comments from Michael
Changes from V4:
- Add ability to negotiate the number of queues through control virtqueue
- Ethtool -{L|l} support and default the tx/rx queue number to 1
- Expose the API to set irq affinity instead of irq itself
Changes from V3:
- Rebase to the net-next
- Let queue 2 to be the control virtqueue to obey the spec
- Prodives irq affinity
- Choose txq based on processor id
Reference:
- Virtio spec RFC: http://patchwork.ozlabs.org/patch/201303/
- V6: https://lkml.org/lkml/2012/10/30/127
- V5: http://lwn.net/Articles/505388/
- V4: https://lkml.org/lkml/2012/6/25/120
- V2: http://lwn.net/Articles/467283/
Perf Numbers:
Will do some basic test and post as a reply to this mail.
Jason Wang (3):
virtio-net: separate fields of sending/receiving queue from
virtnet_info
virtio_net: multiqueue support
virtio-net: change the number of queues through ethtool
drivers/net/virtio_net.c | 716 ++++++++++++++++++++++++++++-----------
include/uapi/linux/virtio_net.h | 16 +
2 files changed, 536 insertions(+), 196 deletions(-)
^ permalink raw reply [flat|nested] 26+ messages in thread
* [net-next rfc v7 1/3] virtio-net: separate fields of sending/receiving queue from virtnet_info
2012-11-27 10:15 [net-next rfc v7 0/3] Multiqueue virtio-net Jason Wang
@ 2012-11-27 10:15 ` Jason Wang
2012-12-03 1:55 ` Rusty Russell
2012-11-27 10:15 ` [net-next rfc v7 2/3] virtio_net: multiqueue support Jason Wang
2012-11-27 10:16 ` [net-next rfc v7 3/3] virtio-net: change the number of queues through ethtool Jason Wang
2 siblings, 1 reply; 26+ messages in thread
From: Jason Wang @ 2012-11-27 10:15 UTC (permalink / raw)
To: rusty, mst, krkumar2, virtualization, netdev, linux-kernel
Cc: kvm, bhutchings, jwhan, shiyer, Jason Wang
To support multiqueue transmitq/receiveq, the first step is to separate queue
related structure from virtnet_info. This patch introduce send_queue and
receive_queue structure and use the pointer to them as the parameter in
functions handling sending/receiving.
Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
drivers/net/virtio_net.c | 289 +++++++++++++++++++++++++---------------------
1 files changed, 158 insertions(+), 131 deletions(-)
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 26c502e..7975133 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -51,16 +51,44 @@ struct virtnet_stats {
u64 rx_packets;
};
+/* Internal representation of a send virtqueue */
+struct send_queue {
+ /* Virtqueue associated with this send _queue */
+ struct virtqueue *vq;
+
+ /* TX: fragments + linear part + virtio header */
+ struct scatterlist sg[MAX_SKB_FRAGS + 2];
+};
+
+/* Internal representation of a receive virtqueue */
+struct receive_queue {
+ /* Virtqueue associated with this receive_queue */
+ struct virtqueue *vq;
+
+ struct napi_struct napi;
+
+ /* Number of input buffers, and max we've ever had. */
+ unsigned int num, max;
+
+ /* Work struct for refilling if we run low on memory. */
+ struct delayed_work refill;
+
+ /* Chain pages by the private ptr. */
+ struct page *pages;
+
+ /* RX: fragments + linear part + virtio header */
+ struct scatterlist sg[MAX_SKB_FRAGS + 2];
+};
+
struct virtnet_info {
struct virtio_device *vdev;
- struct virtqueue *rvq, *svq, *cvq;
+ struct virtqueue *cvq;
struct net_device *dev;
struct napi_struct napi;
+ struct send_queue sq;
+ struct receive_queue rq;
unsigned int status;
- /* Number of input buffers, and max we've ever had. */
- unsigned int num, max;
-
/* I like... big packets and I cannot lie! */
bool big_packets;
@@ -73,21 +101,11 @@ struct virtnet_info {
/* Active statistics */
struct virtnet_stats __percpu *stats;
- /* Work struct for refilling if we run low on memory. */
- struct delayed_work refill;
-
/* Work struct for config space updates */
struct work_struct config_work;
/* Lock for config space updates */
struct mutex config_lock;
-
- /* Chain pages by the private ptr. */
- struct page *pages;
-
- /* fragments + linear part + virtio header */
- struct scatterlist rx_sg[MAX_SKB_FRAGS + 2];
- struct scatterlist tx_sg[MAX_SKB_FRAGS + 2];
};
struct skb_vnet_hdr {
@@ -117,22 +135,22 @@ static inline struct skb_vnet_hdr *skb_vnet_hdr(struct sk_buff *skb)
* private is used to chain pages for big packets, put the whole
* most recent used list in the beginning for reuse
*/
-static void give_pages(struct virtnet_info *vi, struct page *page)
+static void give_pages(struct receive_queue *rq, struct page *page)
{
struct page *end;
- /* Find end of list, sew whole thing into vi->pages. */
+ /* Find end of list, sew whole thing into vi->rq.pages. */
for (end = page; end->private; end = (struct page *)end->private);
- end->private = (unsigned long)vi->pages;
- vi->pages = page;
+ end->private = (unsigned long)rq->pages;
+ rq->pages = page;
}
-static struct page *get_a_page(struct virtnet_info *vi, gfp_t gfp_mask)
+static struct page *get_a_page(struct receive_queue *rq, gfp_t gfp_mask)
{
- struct page *p = vi->pages;
+ struct page *p = rq->pages;
if (p) {
- vi->pages = (struct page *)p->private;
+ rq->pages = (struct page *)p->private;
/* clear private here, it is used to chain pages */
p->private = 0;
} else
@@ -140,12 +158,12 @@ static struct page *get_a_page(struct virtnet_info *vi, gfp_t gfp_mask)
return p;
}
-static void skb_xmit_done(struct virtqueue *svq)
+static void skb_xmit_done(struct virtqueue *vq)
{
- struct virtnet_info *vi = svq->vdev->priv;
+ struct virtnet_info *vi = vq->vdev->priv;
/* Suppress further interrupts. */
- virtqueue_disable_cb(svq);
+ virtqueue_disable_cb(vq);
/* We were probably waiting for more output buffers. */
netif_wake_queue(vi->dev);
@@ -167,9 +185,10 @@ static void set_skb_frag(struct sk_buff *skb, struct page *page,
}
/* Called from bottom half context */
-static struct sk_buff *page_to_skb(struct virtnet_info *vi,
+static struct sk_buff *page_to_skb(struct receive_queue *rq,
struct page *page, unsigned int len)
{
+ struct virtnet_info *vi = rq->vq->vdev->priv;
struct sk_buff *skb;
struct skb_vnet_hdr *hdr;
unsigned int copy, hdr_len, offset;
@@ -224,12 +243,12 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
}
if (page)
- give_pages(vi, page);
+ give_pages(rq, page);
return skb;
}
-static int receive_mergeable(struct virtnet_info *vi, struct sk_buff *skb)
+static int receive_mergeable(struct receive_queue *rq, struct sk_buff *skb)
{
struct skb_vnet_hdr *hdr = skb_vnet_hdr(skb);
struct page *page;
@@ -243,7 +262,7 @@ static int receive_mergeable(struct virtnet_info *vi, struct sk_buff *skb)
skb->dev->stats.rx_length_errors++;
return -EINVAL;
}
- page = virtqueue_get_buf(vi->rvq, &len);
+ page = virtqueue_get_buf(rq->vq, &len);
if (!page) {
pr_debug("%s: rx error: %d buffers missing\n",
skb->dev->name, hdr->mhdr.num_buffers);
@@ -256,14 +275,15 @@ static int receive_mergeable(struct virtnet_info *vi, struct sk_buff *skb)
set_skb_frag(skb, page, 0, &len);
- --vi->num;
+ --rq->num;
}
return 0;
}
-static void receive_buf(struct net_device *dev, void *buf, unsigned int len)
+static void receive_buf(struct receive_queue *rq, void *buf, unsigned int len)
{
- struct virtnet_info *vi = netdev_priv(dev);
+ struct virtnet_info *vi = rq->vq->vdev->priv;
+ struct net_device *dev = vi->dev;
struct virtnet_stats *stats = this_cpu_ptr(vi->stats);
struct sk_buff *skb;
struct page *page;
@@ -273,7 +293,7 @@ static void receive_buf(struct net_device *dev, void *buf, unsigned int len)
pr_debug("%s: short packet %i\n", dev->name, len);
dev->stats.rx_length_errors++;
if (vi->mergeable_rx_bufs || vi->big_packets)
- give_pages(vi, buf);
+ give_pages(rq, buf);
else
dev_kfree_skb(buf);
return;
@@ -285,14 +305,14 @@ static void receive_buf(struct net_device *dev, void *buf, unsigned int len)
skb_trim(skb, len);
} else {
page = buf;
- skb = page_to_skb(vi, page, len);
+ skb = page_to_skb(rq, page, len);
if (unlikely(!skb)) {
dev->stats.rx_dropped++;
- give_pages(vi, page);
+ give_pages(rq, page);
return;
}
if (vi->mergeable_rx_bufs)
- if (receive_mergeable(vi, skb)) {
+ if (receive_mergeable(rq, skb)) {
dev_kfree_skb(skb);
return;
}
@@ -359,8 +379,9 @@ frame_err:
dev_kfree_skb(skb);
}
-static int add_recvbuf_small(struct virtnet_info *vi, gfp_t gfp)
+static int add_recvbuf_small(struct receive_queue *rq, gfp_t gfp)
{
+ struct virtnet_info *vi = rq->vq->vdev->priv;
struct sk_buff *skb;
struct skb_vnet_hdr *hdr;
int err;
@@ -372,77 +393,77 @@ static int add_recvbuf_small(struct virtnet_info *vi, gfp_t gfp)
skb_put(skb, MAX_PACKET_LEN);
hdr = skb_vnet_hdr(skb);
- sg_set_buf(vi->rx_sg, &hdr->hdr, sizeof hdr->hdr);
+ sg_set_buf(rq->sg, &hdr->hdr, sizeof hdr->hdr);
- skb_to_sgvec(skb, vi->rx_sg + 1, 0, skb->len);
+ skb_to_sgvec(skb, rq->sg + 1, 0, skb->len);
- err = virtqueue_add_buf(vi->rvq, vi->rx_sg, 0, 2, skb, gfp);
+ err = virtqueue_add_buf(rq->vq, rq->sg, 0, 2, skb, gfp);
if (err < 0)
dev_kfree_skb(skb);
return err;
}
-static int add_recvbuf_big(struct virtnet_info *vi, gfp_t gfp)
+static int add_recvbuf_big(struct receive_queue *rq, gfp_t gfp)
{
struct page *first, *list = NULL;
char *p;
int i, err, offset;
- /* page in vi->rx_sg[MAX_SKB_FRAGS + 1] is list tail */
+ /* page in rq->sg[MAX_SKB_FRAGS + 1] is list tail */
for (i = MAX_SKB_FRAGS + 1; i > 1; --i) {
- first = get_a_page(vi, gfp);
+ first = get_a_page(rq, gfp);
if (!first) {
if (list)
- give_pages(vi, list);
+ give_pages(rq, list);
return -ENOMEM;
}
- sg_set_buf(&vi->rx_sg[i], page_address(first), PAGE_SIZE);
+ sg_set_buf(&rq->sg[i], page_address(first), PAGE_SIZE);
/* chain new page in list head to match sg */
first->private = (unsigned long)list;
list = first;
}
- first = get_a_page(vi, gfp);
+ first = get_a_page(rq, gfp);
if (!first) {
- give_pages(vi, list);
+ give_pages(rq, list);
return -ENOMEM;
}
p = page_address(first);
- /* vi->rx_sg[0], vi->rx_sg[1] share the same page */
- /* a separated vi->rx_sg[0] for virtio_net_hdr only due to QEMU bug */
- sg_set_buf(&vi->rx_sg[0], p, sizeof(struct virtio_net_hdr));
+ /* rq->sg[0], rq->sg[1] share the same page */
+ /* a separated rq->sg[0] for virtio_net_hdr only due to QEMU bug */
+ sg_set_buf(&rq->sg[0], p, sizeof(struct virtio_net_hdr));
- /* vi->rx_sg[1] for data packet, from offset */
+ /* rq->sg[1] for data packet, from offset */
offset = sizeof(struct padded_vnet_hdr);
- sg_set_buf(&vi->rx_sg[1], p + offset, PAGE_SIZE - offset);
+ sg_set_buf(&rq->sg[1], p + offset, PAGE_SIZE - offset);
/* chain first in list head */
first->private = (unsigned long)list;
- err = virtqueue_add_buf(vi->rvq, vi->rx_sg, 0, MAX_SKB_FRAGS + 2,
+ err = virtqueue_add_buf(rq->vq, rq->sg, 0, MAX_SKB_FRAGS + 2,
first, gfp);
if (err < 0)
- give_pages(vi, first);
+ give_pages(rq, first);
return err;
}
-static int add_recvbuf_mergeable(struct virtnet_info *vi, gfp_t gfp)
+static int add_recvbuf_mergeable(struct receive_queue *rq, gfp_t gfp)
{
struct page *page;
int err;
- page = get_a_page(vi, gfp);
+ page = get_a_page(rq, gfp);
if (!page)
return -ENOMEM;
- sg_init_one(vi->rx_sg, page_address(page), PAGE_SIZE);
+ sg_init_one(rq->sg, page_address(page), PAGE_SIZE);
- err = virtqueue_add_buf(vi->rvq, vi->rx_sg, 0, 1, page, gfp);
+ err = virtqueue_add_buf(rq->vq, rq->sg, 0, 1, page, gfp);
if (err < 0)
- give_pages(vi, page);
+ give_pages(rq, page);
return err;
}
@@ -454,97 +475,101 @@ static int add_recvbuf_mergeable(struct virtnet_info *vi, gfp_t gfp)
* before we're receiving packets, or from refill_work which is
* careful to disable receiving (using napi_disable).
*/
-static bool try_fill_recv(struct virtnet_info *vi, gfp_t gfp)
+static bool try_fill_recv(struct receive_queue *rq, gfp_t gfp)
{
+ struct virtnet_info *vi = rq->vq->vdev->priv;
int err;
bool oom;
do {
if (vi->mergeable_rx_bufs)
- err = add_recvbuf_mergeable(vi, gfp);
+ err = add_recvbuf_mergeable(rq, gfp);
else if (vi->big_packets)
- err = add_recvbuf_big(vi, gfp);
+ err = add_recvbuf_big(rq, gfp);
else
- err = add_recvbuf_small(vi, gfp);
+ err = add_recvbuf_small(rq, gfp);
oom = err == -ENOMEM;
if (err < 0)
break;
- ++vi->num;
+ ++rq->num;
} while (err > 0);
- if (unlikely(vi->num > vi->max))
- vi->max = vi->num;
- virtqueue_kick(vi->rvq);
+ if (unlikely(rq->num > rq->max))
+ rq->max = rq->num;
+ virtqueue_kick(rq->vq);
return !oom;
}
static void skb_recv_done(struct virtqueue *rvq)
{
struct virtnet_info *vi = rvq->vdev->priv;
+ struct receive_queue *rq = &vi->rq;
+
/* Schedule NAPI, Suppress further interrupts if successful. */
- if (napi_schedule_prep(&vi->napi)) {
+ if (napi_schedule_prep(&rq->napi)) {
virtqueue_disable_cb(rvq);
- __napi_schedule(&vi->napi);
+ __napi_schedule(&rq->napi);
}
}
-static void virtnet_napi_enable(struct virtnet_info *vi)
+static void virtnet_napi_enable(struct receive_queue *rq)
{
- napi_enable(&vi->napi);
+ napi_enable(&rq->napi);
/* If all buffers were filled by other side before we napi_enabled, we
* won't get another interrupt, so process any outstanding packets
* now. virtnet_poll wants re-enable the queue, so we disable here.
* We synchronize against interrupts via NAPI_STATE_SCHED */
- if (napi_schedule_prep(&vi->napi)) {
- virtqueue_disable_cb(vi->rvq);
+ if (napi_schedule_prep(&rq->napi)) {
+ virtqueue_disable_cb(rq->vq);
local_bh_disable();
- __napi_schedule(&vi->napi);
+ __napi_schedule(&rq->napi);
local_bh_enable();
}
}
static void refill_work(struct work_struct *work)
{
- struct virtnet_info *vi;
+ struct receive_queue *rq =
+ container_of(work, struct receive_queue, refill.work);
bool still_empty;
- vi = container_of(work, struct virtnet_info, refill.work);
- napi_disable(&vi->napi);
- still_empty = !try_fill_recv(vi, GFP_KERNEL);
- virtnet_napi_enable(vi);
+ napi_disable(&rq->napi);
+ still_empty = !try_fill_recv(rq, GFP_KERNEL);
+ virtnet_napi_enable(rq);
/* In theory, this can happen: if we don't get any buffers in
* we will *never* try to fill again. */
if (still_empty)
- schedule_delayed_work(&vi->refill, HZ/2);
+ schedule_delayed_work(&rq->refill, HZ/2);
}
static int virtnet_poll(struct napi_struct *napi, int budget)
{
- struct virtnet_info *vi = container_of(napi, struct virtnet_info, napi);
+ struct receive_queue *rq =
+ container_of(napi, struct receive_queue, napi);
void *buf;
unsigned int len, received = 0;
again:
while (received < budget &&
- (buf = virtqueue_get_buf(vi->rvq, &len)) != NULL) {
- receive_buf(vi->dev, buf, len);
- --vi->num;
+ (buf = virtqueue_get_buf(rq->vq, &len)) != NULL) {
+ receive_buf(rq, buf, len);
+ --rq->num;
received++;
}
- if (vi->num < vi->max / 2) {
- if (!try_fill_recv(vi, GFP_ATOMIC))
- schedule_delayed_work(&vi->refill, 0);
+ if (rq->num < rq->max / 2) {
+ if (!try_fill_recv(rq, GFP_ATOMIC))
+ schedule_delayed_work(&rq->refill, 0);
}
/* Out of packets? */
if (received < budget) {
napi_complete(napi);
- if (unlikely(!virtqueue_enable_cb(vi->rvq)) &&
+ if (unlikely(!virtqueue_enable_cb(rq->vq)) &&
napi_schedule_prep(napi)) {
- virtqueue_disable_cb(vi->rvq);
+ virtqueue_disable_cb(rq->vq);
__napi_schedule(napi);
goto again;
}
@@ -553,13 +578,14 @@ again:
return received;
}
-static unsigned int free_old_xmit_skbs(struct virtnet_info *vi)
+static unsigned int free_old_xmit_skbs(struct send_queue *sq)
{
struct sk_buff *skb;
unsigned int len, tot_sgs = 0;
+ struct virtnet_info *vi = sq->vq->vdev->priv;
struct virtnet_stats *stats = this_cpu_ptr(vi->stats);
- while ((skb = virtqueue_get_buf(vi->svq, &len)) != NULL) {
+ while ((skb = virtqueue_get_buf(sq->vq, &len)) != NULL) {
pr_debug("Sent skb %p\n", skb);
u64_stats_update_begin(&stats->tx_syncp);
@@ -573,10 +599,11 @@ static unsigned int free_old_xmit_skbs(struct virtnet_info *vi)
return tot_sgs;
}
-static int xmit_skb(struct virtnet_info *vi, struct sk_buff *skb)
+static int xmit_skb(struct send_queue *sq, struct sk_buff *skb)
{
struct skb_vnet_hdr *hdr = skb_vnet_hdr(skb);
const unsigned char *dest = ((struct ethhdr *)skb->data)->h_dest;
+ struct virtnet_info *vi = sq->vq->vdev->priv;
pr_debug("%s: xmit %p %pM\n", vi->dev->name, skb, dest);
@@ -611,25 +638,26 @@ static int xmit_skb(struct virtnet_info *vi, struct sk_buff *skb)
/* Encode metadata header at front. */
if (vi->mergeable_rx_bufs)
- sg_set_buf(vi->tx_sg, &hdr->mhdr, sizeof hdr->mhdr);
+ sg_set_buf(sq->sg, &hdr->mhdr, sizeof hdr->mhdr);
else
- sg_set_buf(vi->tx_sg, &hdr->hdr, sizeof hdr->hdr);
+ sg_set_buf(sq->sg, &hdr->hdr, sizeof hdr->hdr);
- hdr->num_sg = skb_to_sgvec(skb, vi->tx_sg + 1, 0, skb->len) + 1;
- return virtqueue_add_buf(vi->svq, vi->tx_sg, hdr->num_sg,
+ hdr->num_sg = skb_to_sgvec(skb, sq->sg + 1, 0, skb->len) + 1;
+ return virtqueue_add_buf(sq->vq, sq->sg, hdr->num_sg,
0, skb, GFP_ATOMIC);
}
static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
{
struct virtnet_info *vi = netdev_priv(dev);
+ struct send_queue *sq = &vi->sq;
int capacity;
/* Free up any pending old buffers before queueing new ones. */
- free_old_xmit_skbs(vi);
+ free_old_xmit_skbs(sq);
/* Try to transmit */
- capacity = xmit_skb(vi, skb);
+ capacity = xmit_skb(sq, skb);
/* This can happen with OOM and indirect buffers. */
if (unlikely(capacity < 0)) {
@@ -648,7 +676,7 @@ static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
kfree_skb(skb);
return NETDEV_TX_OK;
}
- virtqueue_kick(vi->svq);
+ virtqueue_kick(sq->vq);
/* Don't wait up for transmitted skbs to be freed. */
skb_orphan(skb);
@@ -658,12 +686,12 @@ static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
* before it gets out of hand. Naturally, this wastes entries. */
if (capacity < 2+MAX_SKB_FRAGS) {
netif_stop_queue(dev);
- if (unlikely(!virtqueue_enable_cb_delayed(vi->svq))) {
+ if (unlikely(!virtqueue_enable_cb_delayed(sq->vq))) {
/* More just got used, free them then recheck. */
- capacity += free_old_xmit_skbs(vi);
+ capacity += free_old_xmit_skbs(sq);
if (capacity >= 2+MAX_SKB_FRAGS) {
netif_start_queue(dev);
- virtqueue_disable_cb(vi->svq);
+ virtqueue_disable_cb(sq->vq);
}
}
}
@@ -731,7 +759,7 @@ static void virtnet_netpoll(struct net_device *dev)
{
struct virtnet_info *vi = netdev_priv(dev);
- napi_schedule(&vi->napi);
+ napi_schedule(&vi->rq.napi);
}
#endif
@@ -740,10 +768,10 @@ static int virtnet_open(struct net_device *dev)
struct virtnet_info *vi = netdev_priv(dev);
/* Make sure we have some buffers: if oom use wq. */
- if (!try_fill_recv(vi, GFP_KERNEL))
- schedule_delayed_work(&vi->refill, 0);
+ if (!try_fill_recv(&vi->rq, GFP_KERNEL))
+ schedule_delayed_work(&vi->rq.refill, 0);
- virtnet_napi_enable(vi);
+ virtnet_napi_enable(&vi->rq);
return 0;
}
@@ -807,8 +835,8 @@ static int virtnet_close(struct net_device *dev)
struct virtnet_info *vi = netdev_priv(dev);
/* Make sure refill_work doesn't re-enable napi! */
- cancel_delayed_work_sync(&vi->refill);
- napi_disable(&vi->napi);
+ cancel_delayed_work_sync(&vi->rq.refill);
+ napi_disable(&vi->rq.napi);
return 0;
}
@@ -920,11 +948,10 @@ static void virtnet_get_ringparam(struct net_device *dev,
{
struct virtnet_info *vi = netdev_priv(dev);
- ring->rx_max_pending = virtqueue_get_vring_size(vi->rvq);
- ring->tx_max_pending = virtqueue_get_vring_size(vi->svq);
+ ring->rx_max_pending = virtqueue_get_vring_size(vi->rq.vq);
+ ring->tx_max_pending = virtqueue_get_vring_size(vi->sq.vq);
ring->rx_pending = ring->rx_max_pending;
ring->tx_pending = ring->tx_max_pending;
-
}
@@ -1034,8 +1061,8 @@ static int init_vqs(struct virtnet_info *vi)
if (err)
return err;
- vi->rvq = vqs[0];
- vi->svq = vqs[1];
+ vi->rq.vq = vqs[0];
+ vi->sq.vq = vqs[1];
if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ)) {
vi->cvq = vqs[2];
@@ -1099,22 +1126,22 @@ static int virtnet_probe(struct virtio_device *vdev)
/* Set up our device-specific information */
vi = netdev_priv(dev);
- netif_napi_add(dev, &vi->napi, virtnet_poll, napi_weight);
+ netif_napi_add(dev, &vi->rq.napi, virtnet_poll, napi_weight);
vi->dev = dev;
vi->vdev = vdev;
vdev->priv = vi;
- vi->pages = NULL;
+ vi->rq.pages = NULL;
vi->stats = alloc_percpu(struct virtnet_stats);
err = -ENOMEM;
if (vi->stats == NULL)
goto free;
- INIT_DELAYED_WORK(&vi->refill, refill_work);
+ INIT_DELAYED_WORK(&vi->rq.refill, refill_work);
mutex_init(&vi->config_lock);
vi->config_enable = true;
INIT_WORK(&vi->config_work, virtnet_config_changed_work);
- sg_init_table(vi->rx_sg, ARRAY_SIZE(vi->rx_sg));
- sg_init_table(vi->tx_sg, ARRAY_SIZE(vi->tx_sg));
+ sg_init_table(vi->rq.sg, ARRAY_SIZE(vi->rq.sg));
+ sg_init_table(vi->sq.sg, ARRAY_SIZE(vi->sq.sg));
/* If we can receive ANY GSO packets, we must allocate large ones. */
if (virtio_has_feature(vdev, VIRTIO_NET_F_GUEST_TSO4) ||
@@ -1136,10 +1163,10 @@ static int virtnet_probe(struct virtio_device *vdev)
}
/* Last of all, set up some receive buffers. */
- try_fill_recv(vi, GFP_KERNEL);
+ try_fill_recv(&vi->rq, GFP_KERNEL);
/* If we didn't even get one input buffer, we're useless. */
- if (vi->num == 0) {
+ if (vi->rq.num == 0) {
err = -ENOMEM;
goto unregister;
}
@@ -1172,22 +1199,22 @@ static void free_unused_bufs(struct virtnet_info *vi)
{
void *buf;
while (1) {
- buf = virtqueue_detach_unused_buf(vi->svq);
+ buf = virtqueue_detach_unused_buf(vi->sq.vq);
if (!buf)
break;
dev_kfree_skb(buf);
}
while (1) {
- buf = virtqueue_detach_unused_buf(vi->rvq);
+ buf = virtqueue_detach_unused_buf(vi->rq.vq);
if (!buf)
break;
if (vi->mergeable_rx_bufs || vi->big_packets)
- give_pages(vi, buf);
+ give_pages(&vi->rq, buf);
else
dev_kfree_skb(buf);
- --vi->num;
+ --vi->rq.num;
}
- BUG_ON(vi->num != 0);
+ BUG_ON(vi->rq.num != 0);
}
static void remove_vq_common(struct virtnet_info *vi)
@@ -1199,8 +1226,8 @@ static void remove_vq_common(struct virtnet_info *vi)
vi->vdev->config->del_vqs(vi->vdev);
- while (vi->pages)
- __free_pages(get_a_page(vi, GFP_KERNEL), 0);
+ while (vi->rq.pages)
+ __free_pages(get_a_page(&vi->rq, GFP_KERNEL), 0);
}
static void __devexit virtnet_remove(struct virtio_device *vdev)
@@ -1233,10 +1260,10 @@ static int virtnet_freeze(struct virtio_device *vdev)
mutex_unlock(&vi->config_lock);
netif_device_detach(vi->dev);
- cancel_delayed_work_sync(&vi->refill);
+ cancel_delayed_work_sync(&vi->rq.refill);
if (netif_running(vi->dev))
- napi_disable(&vi->napi);
+ napi_disable(&vi->rq.napi);
remove_vq_common(vi);
@@ -1255,12 +1282,12 @@ static int virtnet_restore(struct virtio_device *vdev)
return err;
if (netif_running(vi->dev))
- virtnet_napi_enable(vi);
+ virtnet_napi_enable(&vi->rq);
netif_device_attach(vi->dev);
- if (!try_fill_recv(vi, GFP_KERNEL))
- schedule_delayed_work(&vi->refill, 0);
+ if (!try_fill_recv(&vi->rq, GFP_KERNEL))
+ schedule_delayed_work(&vi->rq.refill, 0);
mutex_lock(&vi->config_lock);
vi->config_enable = true;
--
1.7.1
^ permalink raw reply related [flat|nested] 26+ messages in thread
* [net-next rfc v7 2/3] virtio_net: multiqueue support
2012-11-27 10:15 [net-next rfc v7 0/3] Multiqueue virtio-net Jason Wang
2012-11-27 10:15 ` [net-next rfc v7 1/3] virtio-net: separate fields of sending/receiving queue from virtnet_info Jason Wang
@ 2012-11-27 10:15 ` Jason Wang
2012-12-02 16:06 ` Michael S. Tsirkin
` (2 more replies)
2012-11-27 10:16 ` [net-next rfc v7 3/3] virtio-net: change the number of queues through ethtool Jason Wang
2 siblings, 3 replies; 26+ messages in thread
From: Jason Wang @ 2012-11-27 10:15 UTC (permalink / raw)
To: rusty, mst, krkumar2, virtualization, netdev, linux-kernel
Cc: kvm, bhutchings, jwhan, shiyer, Jason Wang
This addes multiqueue support to virtio_net driver. In multiple queue modes, the
driver expects the number of queue paris is equal to the number of vcpus. To
eliminate the contention bettwen vcpus and virtqueues, per-cpu virtqueue pairs
were implemented through:
- select the txq based on the smp processor id.
- smp affinity hint were set to the vcpu that owns the queue pairs.
Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
drivers/net/virtio_net.c | 454 ++++++++++++++++++++++++++++++---------
include/uapi/linux/virtio_net.h | 16 ++
2 files changed, 371 insertions(+), 99 deletions(-)
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 7975133..bcaa6e5 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -84,17 +84,25 @@ struct virtnet_info {
struct virtio_device *vdev;
struct virtqueue *cvq;
struct net_device *dev;
- struct napi_struct napi;
- struct send_queue sq;
- struct receive_queue rq;
+ struct send_queue *sq;
+ struct receive_queue *rq;
unsigned int status;
+ /* Max # of queue pairs supported by the device */
+ u16 max_queue_pairs;
+
+ /* # of queue pairs currently used by the driver */
+ u16 curr_queue_pairs;
+
/* I like... big packets and I cannot lie! */
bool big_packets;
/* Host will merge rx buffers for big packets (shake it! shake it!) */
bool mergeable_rx_bufs;
+ /* Has control virtqueue */
+ bool has_cvq;
+
/* enable config space updates */
bool config_enable;
@@ -126,6 +134,34 @@ struct padded_vnet_hdr {
char padding[6];
};
+static const struct ethtool_ops virtnet_ethtool_ops;
+
+/*
+ * Converting between virtqueue no. and kernel tx/rx queue no.
+ * 0:rx0 1:tx0 2:cvq 3:rx1 4:tx1 ... 2N+1:rxN 2N+2:txN
+ */
+static int vq2txq(struct virtqueue *vq)
+{
+ int index = virtqueue_get_queue_index(vq);
+ return index == 1 ? 0 : (index - 2) / 2;
+}
+
+static int txq2vq(int txq)
+{
+ return txq ? 2 * txq + 2 : 1;
+}
+
+static int vq2rxq(struct virtqueue *vq)
+{
+ int index = virtqueue_get_queue_index(vq);
+ return index ? (index - 1) / 2 : 0;
+}
+
+static int rxq2vq(int rxq)
+{
+ return rxq ? 2 * rxq + 1 : 0;
+}
+
static inline struct skb_vnet_hdr *skb_vnet_hdr(struct sk_buff *skb)
{
return (struct skb_vnet_hdr *)skb->cb;
@@ -166,7 +202,7 @@ static void skb_xmit_done(struct virtqueue *vq)
virtqueue_disable_cb(vq);
/* We were probably waiting for more output buffers. */
- netif_wake_queue(vi->dev);
+ netif_wake_subqueue(vi->dev, vq2txq(vq));
}
static void set_skb_frag(struct sk_buff *skb, struct page *page,
@@ -503,7 +539,7 @@ static bool try_fill_recv(struct receive_queue *rq, gfp_t gfp)
static void skb_recv_done(struct virtqueue *rvq)
{
struct virtnet_info *vi = rvq->vdev->priv;
- struct receive_queue *rq = &vi->rq;
+ struct receive_queue *rq = &vi->rq[vq2rxq(rvq)];
/* Schedule NAPI, Suppress further interrupts if successful. */
if (napi_schedule_prep(&rq->napi)) {
@@ -650,7 +686,8 @@ static int xmit_skb(struct send_queue *sq, struct sk_buff *skb)
static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
{
struct virtnet_info *vi = netdev_priv(dev);
- struct send_queue *sq = &vi->sq;
+ int qnum = skb_get_queue_mapping(skb);
+ struct send_queue *sq = &vi->sq[qnum];
int capacity;
/* Free up any pending old buffers before queueing new ones. */
@@ -664,13 +701,14 @@ static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
if (likely(capacity == -ENOMEM)) {
if (net_ratelimit())
dev_warn(&dev->dev,
- "TX queue failure: out of memory\n");
+ "TXQ (%d) failure: out of memory\n",
+ qnum);
} else {
dev->stats.tx_fifo_errors++;
if (net_ratelimit())
dev_warn(&dev->dev,
- "Unexpected TX queue failure: %d\n",
- capacity);
+ "Unexpected TXQ (%d) failure: %d\n",
+ qnum, capacity);
}
dev->stats.tx_dropped++;
kfree_skb(skb);
@@ -685,12 +723,12 @@ static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
/* Apparently nice girls don't return TX_BUSY; stop the queue
* before it gets out of hand. Naturally, this wastes entries. */
if (capacity < 2+MAX_SKB_FRAGS) {
- netif_stop_queue(dev);
+ netif_stop_subqueue(dev, qnum);
if (unlikely(!virtqueue_enable_cb_delayed(sq->vq))) {
/* More just got used, free them then recheck. */
capacity += free_old_xmit_skbs(sq);
if (capacity >= 2+MAX_SKB_FRAGS) {
- netif_start_queue(dev);
+ netif_start_subqueue(dev, qnum);
virtqueue_disable_cb(sq->vq);
}
}
@@ -758,23 +796,13 @@ static struct rtnl_link_stats64 *virtnet_stats(struct net_device *dev,
static void virtnet_netpoll(struct net_device *dev)
{
struct virtnet_info *vi = netdev_priv(dev);
+ int i;
- napi_schedule(&vi->rq.napi);
+ for (i = 0; i < vi->curr_queue_pairs; i++)
+ napi_schedule(&vi->rq[i].napi);
}
#endif
-static int virtnet_open(struct net_device *dev)
-{
- struct virtnet_info *vi = netdev_priv(dev);
-
- /* Make sure we have some buffers: if oom use wq. */
- if (!try_fill_recv(&vi->rq, GFP_KERNEL))
- schedule_delayed_work(&vi->rq.refill, 0);
-
- virtnet_napi_enable(&vi->rq);
- return 0;
-}
-
/*
* Send command via the control virtqueue and check status. Commands
* supported by the hypervisor, as indicated by feature bits, should
@@ -830,13 +858,53 @@ static void virtnet_ack_link_announce(struct virtnet_info *vi)
rtnl_unlock();
}
+static int virtnet_set_queues(struct virtnet_info *vi)
+{
+ struct scatterlist sg;
+ struct virtio_net_ctrl_rfs s;
+ struct net_device *dev = vi->dev;
+
+ s.virtqueue_pairs = vi->curr_queue_pairs;
+ sg_init_one(&sg, &s, sizeof(s));
+
+ if (!vi->has_cvq)
+ return -EINVAL;
+
+ if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RFS,
+ VIRTIO_NET_CTRL_RFS_VQ_PAIRS_SET, &sg, 1, 0)){
+ dev_warn(&dev->dev, "Fail to set the number of queue pairs to"
+ " %d\n", vi->curr_queue_pairs);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+static int virtnet_open(struct net_device *dev)
+{
+ struct virtnet_info *vi = netdev_priv(dev);
+ int i;
+
+ for (i = 0; i < vi->max_queue_pairs; i++) {
+ /* Make sure we have some buffers: if oom use wq. */
+ if (!try_fill_recv(&vi->rq[i], GFP_KERNEL))
+ schedule_delayed_work(&vi->rq[i].refill, 0);
+ virtnet_napi_enable(&vi->rq[i]);
+ }
+
+ return 0;
+}
+
static int virtnet_close(struct net_device *dev)
{
struct virtnet_info *vi = netdev_priv(dev);
+ int i;
/* Make sure refill_work doesn't re-enable napi! */
- cancel_delayed_work_sync(&vi->rq.refill);
- napi_disable(&vi->rq.napi);
+ for (i = 0; i < vi->max_queue_pairs; i++) {
+ cancel_delayed_work_sync(&vi->rq[i].refill);
+ napi_disable(&vi->rq[i].napi);
+ }
return 0;
}
@@ -948,8 +1016,8 @@ static void virtnet_get_ringparam(struct net_device *dev,
{
struct virtnet_info *vi = netdev_priv(dev);
- ring->rx_max_pending = virtqueue_get_vring_size(vi->rq.vq);
- ring->tx_max_pending = virtqueue_get_vring_size(vi->sq.vq);
+ ring->rx_max_pending = virtqueue_get_vring_size(vi->rq[0].vq);
+ ring->tx_max_pending = virtqueue_get_vring_size(vi->sq[0].vq);
ring->rx_pending = ring->rx_max_pending;
ring->tx_pending = ring->tx_max_pending;
}
@@ -967,12 +1035,6 @@ static void virtnet_get_drvinfo(struct net_device *dev,
}
-static const struct ethtool_ops virtnet_ethtool_ops = {
- .get_drvinfo = virtnet_get_drvinfo,
- .get_link = ethtool_op_get_link,
- .get_ringparam = virtnet_get_ringparam,
-};
-
#define MIN_MTU 68
#define MAX_MTU 65535
@@ -984,6 +1046,20 @@ static int virtnet_change_mtu(struct net_device *dev, int new_mtu)
return 0;
}
+/* To avoid contending a lock hold by a vcpu who would exit to host, select the
+ * txq based on the processor id.
+ */
+static u16 virtnet_select_queue(struct net_device *dev, struct sk_buff *skb)
+{
+ int txq = skb_rx_queue_recorded(skb) ? skb_get_rx_queue(skb) :
+ smp_processor_id();
+
+ while (unlikely(txq >= dev->real_num_tx_queues))
+ txq -= dev->real_num_tx_queues;
+
+ return txq;
+}
+
static const struct net_device_ops virtnet_netdev = {
.ndo_open = virtnet_open,
.ndo_stop = virtnet_close,
@@ -995,6 +1071,7 @@ static const struct net_device_ops virtnet_netdev = {
.ndo_get_stats64 = virtnet_stats,
.ndo_vlan_rx_add_vid = virtnet_vlan_rx_add_vid,
.ndo_vlan_rx_kill_vid = virtnet_vlan_rx_kill_vid,
+ .ndo_select_queue = virtnet_select_queue,
#ifdef CONFIG_NET_POLL_CONTROLLER
.ndo_poll_controller = virtnet_netpoll,
#endif
@@ -1030,10 +1107,10 @@ static void virtnet_config_changed_work(struct work_struct *work)
if (vi->status & VIRTIO_NET_S_LINK_UP) {
netif_carrier_on(vi->dev);
- netif_wake_queue(vi->dev);
+ netif_tx_wake_all_queues(vi->dev);
} else {
netif_carrier_off(vi->dev);
- netif_stop_queue(vi->dev);
+ netif_tx_stop_all_queues(vi->dev);
}
done:
mutex_unlock(&vi->config_lock);
@@ -1046,41 +1123,212 @@ static void virtnet_config_changed(struct virtio_device *vdev)
schedule_work(&vi->config_work);
}
-static int init_vqs(struct virtnet_info *vi)
+static void free_receive_bufs(struct virtnet_info *vi)
+{
+ int i;
+
+ for (i = 0; i < vi->max_queue_pairs; i++) {
+ while (vi->rq[i].pages)
+ __free_pages(get_a_page(&vi->rq[i], GFP_KERNEL), 0);
+ }
+}
+
+/* Free memory allocated for send and receive queues */
+static void virtnet_free_queues(struct virtnet_info *vi)
{
- struct virtqueue *vqs[3];
- vq_callback_t *callbacks[] = { skb_recv_done, skb_xmit_done, NULL};
- const char *names[] = { "input", "output", "control" };
- int nvqs, err;
+ kfree(vi->rq);
+ vi->rq = NULL;
+ kfree(vi->sq);
+ vi->sq = NULL;
+}
- /* We expect two virtqueues, receive then send,
- * and optionally control. */
- nvqs = virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ) ? 3 : 2;
+static void free_unused_bufs(struct virtnet_info *vi)
+{
+ void *buf;
+ int i;
- err = vi->vdev->config->find_vqs(vi->vdev, nvqs, vqs, callbacks, names);
- if (err)
- return err;
+ for (i = 0; i < vi->max_queue_pairs; i++) {
+ struct virtqueue *vq = vi->sq[i].vq;
+ while ((buf = virtqueue_detach_unused_buf(vq)) != NULL)
+ dev_kfree_skb(buf);
+ }
- vi->rq.vq = vqs[0];
- vi->sq.vq = vqs[1];
+ for (i = 0; i < vi->max_queue_pairs; i++) {
+ struct virtqueue *vq = vi->rq[i].vq;
- if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ)) {
- vi->cvq = vqs[2];
+ while ((buf = virtqueue_detach_unused_buf(vq)) != NULL) {
+ if (vi->mergeable_rx_bufs || vi->big_packets)
+ give_pages(&vi->rq[i], buf);
+ else
+ dev_kfree_skb(buf);
+ --vi->rq[i].num;
+ }
+ BUG_ON(vi->rq[i].num != 0);
+ }
+}
+static void virtnet_set_affinity(struct virtnet_info *vi, bool set)
+{
+ int i;
+
+ for (i = 0; i < vi->max_queue_pairs; i++) {
+ int cpu = set ? i : -1;
+ virtqueue_set_affinity(vi->rq[i].vq, cpu);
+ virtqueue_set_affinity(vi->sq[i].vq, cpu);
+ }
+}
+
+static void virtnet_del_vqs(struct virtnet_info *vi)
+{
+ struct virtio_device *vdev = vi->vdev;
+
+ virtnet_set_affinity(vi, false);
+
+ vdev->config->del_vqs(vdev);
+
+ virtnet_free_queues(vi);
+}
+
+static int virtnet_find_vqs(struct virtnet_info *vi)
+{
+ vq_callback_t **callbacks;
+ struct virtqueue **vqs;
+ int ret = -ENOMEM;
+ int i, total_vqs;
+ char **names;
+
+ /*
+ * We expect 1 RX virtqueue followed by 1 TX virtqueue, followd by
+ * possible control virtqueue, followed by RX/TX N-1 queue pairs used
+ * in multiqueue mode.
+ */
+ total_vqs = vi->max_queue_pairs * 2 +
+ virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ);
+
+ /* Allocate space for find_vqs parameters */
+ vqs = kzalloc(total_vqs * sizeof(*vqs), GFP_KERNEL);
+ callbacks = kzalloc(total_vqs * sizeof(*callbacks), GFP_KERNEL);
+ if (!vqs || !callbacks)
+ goto err_mem;
+ names = kzalloc(total_vqs * sizeof(*names), GFP_KERNEL);
+ if (!names)
+ goto err_mem;
+
+ /* Parameters for control virtqueue, if any */
+ if (vi->has_cvq) {
+ callbacks[2] = NULL;
+ names[2] = kasprintf(GFP_KERNEL, "control");
+ }
+
+ /* Allocate/initialize parameters for send/receive virtqueues */
+ for (i = 0; i < vi->max_queue_pairs; i++) {
+ callbacks[rxq2vq(i)] = skb_recv_done;
+ callbacks[txq2vq(i)] = skb_xmit_done;
+ names[rxq2vq(i)] = kasprintf(GFP_KERNEL, "input.%d", i);
+ names[txq2vq(i)] = kasprintf(GFP_KERNEL, "output.%d", i);
+ }
+
+ ret = vi->vdev->config->find_vqs(vi->vdev, total_vqs, vqs, callbacks,
+ (const char **)names);
+ if (ret)
+ goto err_names;
+
+ if (vi->has_cvq) {
+ vi->cvq = vqs[2];
if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VLAN))
vi->dev->features |= NETIF_F_HW_VLAN_FILTER;
}
+
+ for (i = 0; i < vi->max_queue_pairs; i++) {
+ vi->rq[i].vq = vqs[rxq2vq(i)];
+ vi->sq[i].vq = vqs[txq2vq(i)];
+ }
+
+ kfree(callbacks);
+ kfree(vqs);
+
+ return 0;
+
+err_names:
+ for (i = 0; i < total_vqs * 2; i ++)
+ kfree(names[i]);
+ kfree(names);
+
+err_mem:
+ kfree(callbacks);
+ kfree(vqs);
+
+ return ret;
+}
+
+static int virtnet_alloc_queues(struct virtnet_info *vi)
+{
+ int i;
+
+ vi->sq = kzalloc(sizeof(vi->sq[0]) * vi->max_queue_pairs, GFP_KERNEL);
+ vi->rq = kzalloc(sizeof(vi->rq[0]) * vi->max_queue_pairs, GFP_KERNEL);
+ if (!vi->rq || !vi->sq)
+ goto err;
+
+ /* setup initial receive and send queue parameters */
+ for (i = 0; i < vi->max_queue_pairs; i++) {
+ vi->rq[i].pages = NULL;
+ INIT_DELAYED_WORK(&vi->rq[i].refill, refill_work);
+ netif_napi_add(vi->dev, &vi->rq[i].napi, virtnet_poll,
+ napi_weight);
+
+ sg_init_table(vi->rq[i].sg, ARRAY_SIZE(vi->rq[i].sg));
+ sg_init_table(vi->sq[i].sg, ARRAY_SIZE(vi->sq[i].sg));
+ }
+
+
return 0;
+
+err:
+ virtnet_free_queues(vi);
+ return -ENOMEM;
+}
+
+static int init_vqs(struct virtnet_info *vi)
+{
+ int ret;
+
+ /* Allocate send & receive queues */
+ ret = virtnet_alloc_queues(vi);
+ if (ret)
+ goto err;
+
+ ret = virtnet_find_vqs(vi);
+ if (ret)
+ goto err_free;
+
+ virtnet_set_affinity(vi, true);
+ return 0;
+
+err_free:
+ virtnet_free_queues(vi);
+err:
+ return ret;
}
static int virtnet_probe(struct virtio_device *vdev)
{
- int err;
+ int i, err;
struct net_device *dev;
struct virtnet_info *vi;
+ u16 curr_queue_pairs;
+
+ /* Find if host supports multiqueue virtio_net device */
+ err = virtio_config_val(vdev, VIRTIO_NET_F_RFS,
+ offsetof(struct virtio_net_config,
+ max_virtqueue_pairs), &curr_queue_pairs);
+
+ /* We need at least 2 queue's */
+ if (err)
+ curr_queue_pairs = 1;
/* Allocate ourselves a network device with room for our info */
- dev = alloc_etherdev(sizeof(struct virtnet_info));
+ dev = alloc_etherdev_mq(sizeof(struct virtnet_info), curr_queue_pairs);
if (!dev)
return -ENOMEM;
@@ -1126,22 +1374,17 @@ static int virtnet_probe(struct virtio_device *vdev)
/* Set up our device-specific information */
vi = netdev_priv(dev);
- netif_napi_add(dev, &vi->rq.napi, virtnet_poll, napi_weight);
vi->dev = dev;
vi->vdev = vdev;
vdev->priv = vi;
- vi->rq.pages = NULL;
vi->stats = alloc_percpu(struct virtnet_stats);
err = -ENOMEM;
if (vi->stats == NULL)
goto free;
- INIT_DELAYED_WORK(&vi->rq.refill, refill_work);
mutex_init(&vi->config_lock);
vi->config_enable = true;
INIT_WORK(&vi->config_work, virtnet_config_changed_work);
- sg_init_table(vi->rq.sg, ARRAY_SIZE(vi->rq.sg));
- sg_init_table(vi->sq.sg, ARRAY_SIZE(vi->sq.sg));
/* If we can receive ANY GSO packets, we must allocate large ones. */
if (virtio_has_feature(vdev, VIRTIO_NET_F_GUEST_TSO4) ||
@@ -1152,10 +1395,21 @@ static int virtnet_probe(struct virtio_device *vdev)
if (virtio_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF))
vi->mergeable_rx_bufs = true;
+ if (virtio_has_feature(vdev, VIRTIO_NET_F_CTRL_VQ))
+ vi->has_cvq = true;
+
+ /* Use single tx/rx queue pair as default */
+ vi->curr_queue_pairs = 1;
+ vi->max_queue_pairs = curr_queue_pairs;
+
+ /* Allocate/initialize the rx/tx queues, and invoke find_vqs */
err = init_vqs(vi);
if (err)
goto free_stats;
+ netif_set_real_num_tx_queues(dev, 1);
+ netif_set_real_num_rx_queues(dev, 1);
+
err = register_netdev(dev);
if (err) {
pr_debug("virtio_net: registering device failed\n");
@@ -1163,12 +1417,15 @@ static int virtnet_probe(struct virtio_device *vdev)
}
/* Last of all, set up some receive buffers. */
- try_fill_recv(&vi->rq, GFP_KERNEL);
-
- /* If we didn't even get one input buffer, we're useless. */
- if (vi->rq.num == 0) {
- err = -ENOMEM;
- goto unregister;
+ for (i = 0; i < vi->max_queue_pairs; i++) {
+ try_fill_recv(&vi->rq[i], GFP_KERNEL);
+
+ /* If we didn't even get one input buffer, we're useless. */
+ if (vi->rq[i].num == 0) {
+ free_unused_bufs(vi);
+ err = -ENOMEM;
+ goto free_recv_bufs;
+ }
}
/* Assume link up if device can't report link status,
@@ -1181,13 +1438,20 @@ static int virtnet_probe(struct virtio_device *vdev)
netif_carrier_on(dev);
}
- pr_debug("virtnet: registered device %s\n", dev->name);
+ pr_debug("virtnet: registered device %s with %d RX and TX vq's\n",
+ dev->name, curr_queue_pairs);
+
return 0;
-unregister:
+free_recv_bufs:
+ free_receive_bufs(vi);
unregister_netdev(dev);
+
free_vqs:
- vdev->config->del_vqs(vdev);
+ for (i = 0; i <curr_queue_pairs; i++)
+ cancel_delayed_work_sync(&vi->rq[i].refill);
+ virtnet_del_vqs(vi);
+
free_stats:
free_percpu(vi->stats);
free:
@@ -1195,28 +1459,6 @@ free:
return err;
}
-static void free_unused_bufs(struct virtnet_info *vi)
-{
- void *buf;
- while (1) {
- buf = virtqueue_detach_unused_buf(vi->sq.vq);
- if (!buf)
- break;
- dev_kfree_skb(buf);
- }
- while (1) {
- buf = virtqueue_detach_unused_buf(vi->rq.vq);
- if (!buf)
- break;
- if (vi->mergeable_rx_bufs || vi->big_packets)
- give_pages(&vi->rq, buf);
- else
- dev_kfree_skb(buf);
- --vi->rq.num;
- }
- BUG_ON(vi->rq.num != 0);
-}
-
static void remove_vq_common(struct virtnet_info *vi)
{
vi->vdev->config->reset(vi->vdev);
@@ -1224,10 +1466,9 @@ static void remove_vq_common(struct virtnet_info *vi)
/* Free unused buffers in both send and recv, if any. */
free_unused_bufs(vi);
- vi->vdev->config->del_vqs(vi->vdev);
+ free_receive_bufs(vi);
- while (vi->rq.pages)
- __free_pages(get_a_page(&vi->rq, GFP_KERNEL), 0);
+ virtnet_del_vqs(vi);
}
static void __devexit virtnet_remove(struct virtio_device *vdev)
@@ -1253,6 +1494,7 @@ static void __devexit virtnet_remove(struct virtio_device *vdev)
static int virtnet_freeze(struct virtio_device *vdev)
{
struct virtnet_info *vi = vdev->priv;
+ int i;
/* Prevent config work handler from accessing the device */
mutex_lock(&vi->config_lock);
@@ -1260,10 +1502,14 @@ static int virtnet_freeze(struct virtio_device *vdev)
mutex_unlock(&vi->config_lock);
netif_device_detach(vi->dev);
- cancel_delayed_work_sync(&vi->rq.refill);
+ for (i = 0; i < vi->max_queue_pairs; i++)
+ cancel_delayed_work_sync(&vi->rq[i].refill);
if (netif_running(vi->dev))
- napi_disable(&vi->rq.napi);
+ for (i = 0; i < vi->max_queue_pairs; i++) {
+ napi_disable(&vi->rq[i].napi);
+ netif_napi_del(&vi->rq[i].napi);
+ }
remove_vq_common(vi);
@@ -1275,24 +1521,28 @@ static int virtnet_freeze(struct virtio_device *vdev)
static int virtnet_restore(struct virtio_device *vdev)
{
struct virtnet_info *vi = vdev->priv;
- int err;
+ int err, i;
err = init_vqs(vi);
if (err)
return err;
if (netif_running(vi->dev))
- virtnet_napi_enable(&vi->rq);
+ for (i = 0; i < vi->max_queue_pairs; i++)
+ virtnet_napi_enable(&vi->rq[i]);
netif_device_attach(vi->dev);
- if (!try_fill_recv(&vi->rq, GFP_KERNEL))
- schedule_delayed_work(&vi->rq.refill, 0);
+ for (i = 0; i < vi->max_queue_pairs; i++)
+ if (!try_fill_recv(&vi->rq[i], GFP_KERNEL))
+ schedule_delayed_work(&vi->rq[i].refill, 0);
mutex_lock(&vi->config_lock);
vi->config_enable = true;
mutex_unlock(&vi->config_lock);
+ BUG_ON(virtnet_set_queues(vi));
+
return 0;
}
#endif
@@ -1310,7 +1560,7 @@ static unsigned int features[] = {
VIRTIO_NET_F_GUEST_ECN, VIRTIO_NET_F_GUEST_UFO,
VIRTIO_NET_F_MRG_RXBUF, VIRTIO_NET_F_STATUS, VIRTIO_NET_F_CTRL_VQ,
VIRTIO_NET_F_CTRL_RX, VIRTIO_NET_F_CTRL_VLAN,
- VIRTIO_NET_F_GUEST_ANNOUNCE,
+ VIRTIO_NET_F_GUEST_ANNOUNCE, VIRTIO_NET_F_RFS,
};
static struct virtio_driver virtio_net_driver = {
@@ -1328,6 +1578,12 @@ static struct virtio_driver virtio_net_driver = {
#endif
};
+static const struct ethtool_ops virtnet_ethtool_ops = {
+ .get_drvinfo = virtnet_get_drvinfo,
+ .get_link = ethtool_op_get_link,
+ .get_ringparam = virtnet_get_ringparam,
+};
+
static int __init init(void)
{
return register_virtio_driver(&virtio_net_driver);
diff --git a/include/uapi/linux/virtio_net.h b/include/uapi/linux/virtio_net.h
index 2470f54..6056cec 100644
--- a/include/uapi/linux/virtio_net.h
+++ b/include/uapi/linux/virtio_net.h
@@ -51,6 +51,7 @@
#define VIRTIO_NET_F_CTRL_RX_EXTRA 20 /* Extra RX mode control support */
#define VIRTIO_NET_F_GUEST_ANNOUNCE 21 /* Guest can announce device on the
* network */
+#define VIRTIO_NET_F_RFS 22 /* Device supports multiple TXQ/RXQ */
#define VIRTIO_NET_S_LINK_UP 1 /* Link is up */
#define VIRTIO_NET_S_ANNOUNCE 2 /* Announcement is needed */
@@ -60,6 +61,8 @@ struct virtio_net_config {
__u8 mac[6];
/* See VIRTIO_NET_F_STATUS and VIRTIO_NET_S_* above */
__u16 status;
+ /* Total number of RX/TX queues */
+ __u16 max_virtqueue_pairs;
} __attribute__((packed));
/* This is the first element of the scatter-gather list. If you don't
@@ -166,4 +169,17 @@ struct virtio_net_ctrl_mac {
#define VIRTIO_NET_CTRL_ANNOUNCE 3
#define VIRTIO_NET_CTRL_ANNOUNCE_ACK 0
+/*
+ * Control multiqueue
+ *
+ */
+struct virtio_net_ctrl_rfs {
+ u16 virtqueue_pairs;
+};
+
+#define VIRTIO_NET_CTRL_RFS 4
+ #define VIRTIO_NET_CTRL_RFS_VQ_PAIRS_SET 0
+ #define VIRTIO_NET_CTRL_RFS_VQ_PAIRS_MIN 1
+ #define VIRTIO_NET_CTRL_RFS_VQ_PAIRS_MAX 0x8000
+
#endif /* _LINUX_VIRTIO_NET_H */
--
1.7.1
^ permalink raw reply related [flat|nested] 26+ messages in thread
* [net-next rfc v7 3/3] virtio-net: change the number of queues through ethtool
2012-11-27 10:15 [net-next rfc v7 0/3] Multiqueue virtio-net Jason Wang
2012-11-27 10:15 ` [net-next rfc v7 1/3] virtio-net: separate fields of sending/receiving queue from virtnet_info Jason Wang
2012-11-27 10:15 ` [net-next rfc v7 2/3] virtio_net: multiqueue support Jason Wang
@ 2012-11-27 10:16 ` Jason Wang
2012-12-02 16:09 ` Michael S. Tsirkin
2 siblings, 1 reply; 26+ messages in thread
From: Jason Wang @ 2012-11-27 10:16 UTC (permalink / raw)
To: rusty, mst, krkumar2, virtualization, netdev, linux-kernel
Cc: kvm, bhutchings, jwhan, shiyer, Jason Wang
This patch implement the {set|get}_channels method of ethool to allow user to
change the number of queues dymaically when the device is running. This would
let the user to configure it on demand.
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
drivers/net/virtio_net.c | 41 +++++++++++++++++++++++++++++++++++++++++
1 files changed, 41 insertions(+), 0 deletions(-)
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index bcaa6e5..f08ec2a 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -1578,10 +1578,51 @@ static struct virtio_driver virtio_net_driver = {
#endif
};
+/* TODO: Eliminate OOO packets during switching */
+static int virtnet_set_channels(struct net_device *dev,
+ struct ethtool_channels *channels)
+{
+ struct virtnet_info *vi = netdev_priv(dev);
+ u16 queue_pairs = channels->combined_count;
+
+ /* We don't support separate rx/tx channels.
+ * We don't allow setting 'other' channels.
+ */
+ if (channels->rx_count || channels->tx_count || channels->other_count)
+ return -EINVAL;
+
+ /* Only two modes were support currently */
+ if (queue_pairs != vi->max_queue_pairs && queue_pairs != 1)
+ return -EINVAL;
+
+ vi->curr_queue_pairs = queue_pairs;
+ BUG_ON(virtnet_set_queues(vi));
+
+ netif_set_real_num_tx_queues(dev, vi->curr_queue_pairs);
+ netif_set_real_num_rx_queues(dev, vi->curr_queue_pairs);
+
+ return 0;
+}
+
+static void virtnet_get_channels(struct net_device *dev,
+ struct ethtool_channels *channels)
+{
+ struct virtnet_info *vi = netdev_priv(dev);
+
+ channels->combined_count = vi->curr_queue_pairs;
+ channels->max_combined = vi->max_queue_pairs;
+ channels->max_other = 0;
+ channels->rx_count = 0;
+ channels->tx_count = 0;
+ channels->other_count = 0;
+}
+
static const struct ethtool_ops virtnet_ethtool_ops = {
.get_drvinfo = virtnet_get_drvinfo,
.get_link = ethtool_op_get_link,
.get_ringparam = virtnet_get_ringparam,
+ .set_channels = virtnet_set_channels,
+ .get_channels = virtnet_get_channels,
};
static int __init init(void)
--
1.7.1
^ permalink raw reply related [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 2/3] virtio_net: multiqueue support
2012-11-27 10:15 ` [net-next rfc v7 2/3] virtio_net: multiqueue support Jason Wang
@ 2012-12-02 16:06 ` Michael S. Tsirkin
2012-12-03 5:47 ` Jason Wang
2012-12-03 2:04 ` Rusty Russell
2012-12-03 10:14 ` Michael S. Tsirkin
2 siblings, 1 reply; 26+ messages in thread
From: Michael S. Tsirkin @ 2012-12-02 16:06 UTC (permalink / raw)
To: Jason Wang
Cc: rusty, krkumar2, virtualization, netdev, linux-kernel, kvm,
bhutchings, jwhan, shiyer
On Tue, Nov 27, 2012 at 06:15:59PM +0800, Jason Wang wrote:
> This addes multiqueue support to virtio_net driver. In multiple queue modes, the
> driver expects the number of queue paris is equal to the number of vcpus. To
> eliminate the contention bettwen vcpus and virtqueues, per-cpu virtqueue pairs
> were implemented through:
>
> - select the txq based on the smp processor id.
> - smp affinity hint were set to the vcpu that owns the queue pairs.
>
> Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>
> Signed-off-by: Jason Wang <jasowang@redhat.com>
> ---
> drivers/net/virtio_net.c | 454 ++++++++++++++++++++++++++++++---------
> include/uapi/linux/virtio_net.h | 16 ++
> 2 files changed, 371 insertions(+), 99 deletions(-)
>
> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> index 7975133..bcaa6e5 100644
> --- a/drivers/net/virtio_net.c
> +++ b/drivers/net/virtio_net.c
> @@ -84,17 +84,25 @@ struct virtnet_info {
> struct virtio_device *vdev;
> struct virtqueue *cvq;
> struct net_device *dev;
> - struct napi_struct napi;
> - struct send_queue sq;
> - struct receive_queue rq;
> + struct send_queue *sq;
> + struct receive_queue *rq;
> unsigned int status;
>
> + /* Max # of queue pairs supported by the device */
> + u16 max_queue_pairs;
> +
> + /* # of queue pairs currently used by the driver */
> + u16 curr_queue_pairs;
> +
> /* I like... big packets and I cannot lie! */
> bool big_packets;
>
> /* Host will merge rx buffers for big packets (shake it! shake it!) */
> bool mergeable_rx_bufs;
>
> + /* Has control virtqueue */
> + bool has_cvq;
> +
> /* enable config space updates */
> bool config_enable;
>
> @@ -126,6 +134,34 @@ struct padded_vnet_hdr {
> char padding[6];
> };
>
> +static const struct ethtool_ops virtnet_ethtool_ops;
> +
> +/*
> + * Converting between virtqueue no. and kernel tx/rx queue no.
> + * 0:rx0 1:tx0 2:cvq 3:rx1 4:tx1 ... 2N+1:rxN 2N+2:txN
> + */
Weird, this isn't what spec v5 says: it says
0:rx0 1:tx0 2: rx1 3: tx1 .... vcq
We can change the spec to match but keeping all rx/tx
together seems a bit prettier?
> +static int vq2txq(struct virtqueue *vq)
> +{
> + int index = virtqueue_get_queue_index(vq);
> + return index == 1 ? 0 : (index - 2) / 2;
> +}
> +
> +static int txq2vq(int txq)
> +{
> + return txq ? 2 * txq + 2 : 1;
> +}
> +
> +static int vq2rxq(struct virtqueue *vq)
> +{
> + int index = virtqueue_get_queue_index(vq);
> + return index ? (index - 1) / 2 : 0;
> +}
> +
> +static int rxq2vq(int rxq)
> +{
> + return rxq ? 2 * rxq + 1 : 0;
> +}
> +
> static inline struct skb_vnet_hdr *skb_vnet_hdr(struct sk_buff *skb)
> {
> return (struct skb_vnet_hdr *)skb->cb;
> @@ -166,7 +202,7 @@ static void skb_xmit_done(struct virtqueue *vq)
> virtqueue_disable_cb(vq);
>
> /* We were probably waiting for more output buffers. */
> - netif_wake_queue(vi->dev);
> + netif_wake_subqueue(vi->dev, vq2txq(vq));
> }
>
> static void set_skb_frag(struct sk_buff *skb, struct page *page,
> @@ -503,7 +539,7 @@ static bool try_fill_recv(struct receive_queue *rq, gfp_t gfp)
> static void skb_recv_done(struct virtqueue *rvq)
> {
> struct virtnet_info *vi = rvq->vdev->priv;
> - struct receive_queue *rq = &vi->rq;
> + struct receive_queue *rq = &vi->rq[vq2rxq(rvq)];
>
> /* Schedule NAPI, Suppress further interrupts if successful. */
> if (napi_schedule_prep(&rq->napi)) {
> @@ -650,7 +686,8 @@ static int xmit_skb(struct send_queue *sq, struct sk_buff *skb)
> static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
> {
> struct virtnet_info *vi = netdev_priv(dev);
> - struct send_queue *sq = &vi->sq;
> + int qnum = skb_get_queue_mapping(skb);
> + struct send_queue *sq = &vi->sq[qnum];
> int capacity;
>
> /* Free up any pending old buffers before queueing new ones. */
> @@ -664,13 +701,14 @@ static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
> if (likely(capacity == -ENOMEM)) {
> if (net_ratelimit())
> dev_warn(&dev->dev,
> - "TX queue failure: out of memory\n");
> + "TXQ (%d) failure: out of memory\n",
> + qnum);
> } else {
> dev->stats.tx_fifo_errors++;
> if (net_ratelimit())
> dev_warn(&dev->dev,
> - "Unexpected TX queue failure: %d\n",
> - capacity);
> + "Unexpected TXQ (%d) failure: %d\n",
> + qnum, capacity);
> }
> dev->stats.tx_dropped++;
> kfree_skb(skb);
> @@ -685,12 +723,12 @@ static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
> /* Apparently nice girls don't return TX_BUSY; stop the queue
> * before it gets out of hand. Naturally, this wastes entries. */
> if (capacity < 2+MAX_SKB_FRAGS) {
> - netif_stop_queue(dev);
> + netif_stop_subqueue(dev, qnum);
> if (unlikely(!virtqueue_enable_cb_delayed(sq->vq))) {
> /* More just got used, free them then recheck. */
> capacity += free_old_xmit_skbs(sq);
> if (capacity >= 2+MAX_SKB_FRAGS) {
> - netif_start_queue(dev);
> + netif_start_subqueue(dev, qnum);
> virtqueue_disable_cb(sq->vq);
> }
> }
> @@ -758,23 +796,13 @@ static struct rtnl_link_stats64 *virtnet_stats(struct net_device *dev,
> static void virtnet_netpoll(struct net_device *dev)
> {
> struct virtnet_info *vi = netdev_priv(dev);
> + int i;
>
> - napi_schedule(&vi->rq.napi);
> + for (i = 0; i < vi->curr_queue_pairs; i++)
> + napi_schedule(&vi->rq[i].napi);
> }
> #endif
>
> -static int virtnet_open(struct net_device *dev)
> -{
> - struct virtnet_info *vi = netdev_priv(dev);
> -
> - /* Make sure we have some buffers: if oom use wq. */
> - if (!try_fill_recv(&vi->rq, GFP_KERNEL))
> - schedule_delayed_work(&vi->rq.refill, 0);
> -
> - virtnet_napi_enable(&vi->rq);
> - return 0;
> -}
> -
> /*
> * Send command via the control virtqueue and check status. Commands
> * supported by the hypervisor, as indicated by feature bits, should
> @@ -830,13 +858,53 @@ static void virtnet_ack_link_announce(struct virtnet_info *vi)
> rtnl_unlock();
> }
>
> +static int virtnet_set_queues(struct virtnet_info *vi)
> +{
> + struct scatterlist sg;
> + struct virtio_net_ctrl_rfs s;
> + struct net_device *dev = vi->dev;
> +
> + s.virtqueue_pairs = vi->curr_queue_pairs;
> + sg_init_one(&sg, &s, sizeof(s));
> +
> + if (!vi->has_cvq)
> + return -EINVAL;
> +
> + if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RFS,
> + VIRTIO_NET_CTRL_RFS_VQ_PAIRS_SET, &sg, 1, 0)){
> + dev_warn(&dev->dev, "Fail to set the number of queue pairs to"
> + " %d\n", vi->curr_queue_pairs);
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
> +
> +static int virtnet_open(struct net_device *dev)
> +{
> + struct virtnet_info *vi = netdev_priv(dev);
> + int i;
> +
> + for (i = 0; i < vi->max_queue_pairs; i++) {
> + /* Make sure we have some buffers: if oom use wq. */
> + if (!try_fill_recv(&vi->rq[i], GFP_KERNEL))
> + schedule_delayed_work(&vi->rq[i].refill, 0);
> + virtnet_napi_enable(&vi->rq[i]);
> + }
> +
> + return 0;
> +}
> +
> static int virtnet_close(struct net_device *dev)
> {
> struct virtnet_info *vi = netdev_priv(dev);
> + int i;
>
> /* Make sure refill_work doesn't re-enable napi! */
> - cancel_delayed_work_sync(&vi->rq.refill);
> - napi_disable(&vi->rq.napi);
> + for (i = 0; i < vi->max_queue_pairs; i++) {
> + cancel_delayed_work_sync(&vi->rq[i].refill);
> + napi_disable(&vi->rq[i].napi);
> + }
>
> return 0;
> }
> @@ -948,8 +1016,8 @@ static void virtnet_get_ringparam(struct net_device *dev,
> {
> struct virtnet_info *vi = netdev_priv(dev);
>
> - ring->rx_max_pending = virtqueue_get_vring_size(vi->rq.vq);
> - ring->tx_max_pending = virtqueue_get_vring_size(vi->sq.vq);
> + ring->rx_max_pending = virtqueue_get_vring_size(vi->rq[0].vq);
> + ring->tx_max_pending = virtqueue_get_vring_size(vi->sq[0].vq);
> ring->rx_pending = ring->rx_max_pending;
> ring->tx_pending = ring->tx_max_pending;
> }
> @@ -967,12 +1035,6 @@ static void virtnet_get_drvinfo(struct net_device *dev,
>
> }
>
> -static const struct ethtool_ops virtnet_ethtool_ops = {
> - .get_drvinfo = virtnet_get_drvinfo,
> - .get_link = ethtool_op_get_link,
> - .get_ringparam = virtnet_get_ringparam,
> -};
> -
> #define MIN_MTU 68
> #define MAX_MTU 65535
>
> @@ -984,6 +1046,20 @@ static int virtnet_change_mtu(struct net_device *dev, int new_mtu)
> return 0;
> }
>
> +/* To avoid contending a lock hold by a vcpu who would exit to host, select the
> + * txq based on the processor id.
> + */
> +static u16 virtnet_select_queue(struct net_device *dev, struct sk_buff *skb)
> +{
> + int txq = skb_rx_queue_recorded(skb) ? skb_get_rx_queue(skb) :
> + smp_processor_id();
> +
> + while (unlikely(txq >= dev->real_num_tx_queues))
> + txq -= dev->real_num_tx_queues;
> +
> + return txq;
> +}
> +
> static const struct net_device_ops virtnet_netdev = {
> .ndo_open = virtnet_open,
> .ndo_stop = virtnet_close,
> @@ -995,6 +1071,7 @@ static const struct net_device_ops virtnet_netdev = {
> .ndo_get_stats64 = virtnet_stats,
> .ndo_vlan_rx_add_vid = virtnet_vlan_rx_add_vid,
> .ndo_vlan_rx_kill_vid = virtnet_vlan_rx_kill_vid,
> + .ndo_select_queue = virtnet_select_queue,
> #ifdef CONFIG_NET_POLL_CONTROLLER
> .ndo_poll_controller = virtnet_netpoll,
> #endif
> @@ -1030,10 +1107,10 @@ static void virtnet_config_changed_work(struct work_struct *work)
>
> if (vi->status & VIRTIO_NET_S_LINK_UP) {
> netif_carrier_on(vi->dev);
> - netif_wake_queue(vi->dev);
> + netif_tx_wake_all_queues(vi->dev);
> } else {
> netif_carrier_off(vi->dev);
> - netif_stop_queue(vi->dev);
> + netif_tx_stop_all_queues(vi->dev);
> }
> done:
> mutex_unlock(&vi->config_lock);
> @@ -1046,41 +1123,212 @@ static void virtnet_config_changed(struct virtio_device *vdev)
> schedule_work(&vi->config_work);
> }
>
> -static int init_vqs(struct virtnet_info *vi)
> +static void free_receive_bufs(struct virtnet_info *vi)
> +{
> + int i;
> +
> + for (i = 0; i < vi->max_queue_pairs; i++) {
> + while (vi->rq[i].pages)
> + __free_pages(get_a_page(&vi->rq[i], GFP_KERNEL), 0);
> + }
> +}
> +
> +/* Free memory allocated for send and receive queues */
> +static void virtnet_free_queues(struct virtnet_info *vi)
> {
> - struct virtqueue *vqs[3];
> - vq_callback_t *callbacks[] = { skb_recv_done, skb_xmit_done, NULL};
> - const char *names[] = { "input", "output", "control" };
> - int nvqs, err;
> + kfree(vi->rq);
> + vi->rq = NULL;
> + kfree(vi->sq);
> + vi->sq = NULL;
> +}
>
> - /* We expect two virtqueues, receive then send,
> - * and optionally control. */
> - nvqs = virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ) ? 3 : 2;
> +static void free_unused_bufs(struct virtnet_info *vi)
> +{
> + void *buf;
> + int i;
>
> - err = vi->vdev->config->find_vqs(vi->vdev, nvqs, vqs, callbacks, names);
> - if (err)
> - return err;
> + for (i = 0; i < vi->max_queue_pairs; i++) {
> + struct virtqueue *vq = vi->sq[i].vq;
> + while ((buf = virtqueue_detach_unused_buf(vq)) != NULL)
> + dev_kfree_skb(buf);
> + }
>
> - vi->rq.vq = vqs[0];
> - vi->sq.vq = vqs[1];
> + for (i = 0; i < vi->max_queue_pairs; i++) {
> + struct virtqueue *vq = vi->rq[i].vq;
>
> - if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ)) {
> - vi->cvq = vqs[2];
> + while ((buf = virtqueue_detach_unused_buf(vq)) != NULL) {
> + if (vi->mergeable_rx_bufs || vi->big_packets)
> + give_pages(&vi->rq[i], buf);
> + else
> + dev_kfree_skb(buf);
> + --vi->rq[i].num;
> + }
> + BUG_ON(vi->rq[i].num != 0);
> + }
> +}
>
> +static void virtnet_set_affinity(struct virtnet_info *vi, bool set)
> +{
> + int i;
> +
> + for (i = 0; i < vi->max_queue_pairs; i++) {
> + int cpu = set ? i : -1;
> + virtqueue_set_affinity(vi->rq[i].vq, cpu);
> + virtqueue_set_affinity(vi->sq[i].vq, cpu);
> + }
> +}
> +
> +static void virtnet_del_vqs(struct virtnet_info *vi)
> +{
> + struct virtio_device *vdev = vi->vdev;
> +
> + virtnet_set_affinity(vi, false);
> +
> + vdev->config->del_vqs(vdev);
> +
> + virtnet_free_queues(vi);
> +}
> +
> +static int virtnet_find_vqs(struct virtnet_info *vi)
> +{
> + vq_callback_t **callbacks;
> + struct virtqueue **vqs;
> + int ret = -ENOMEM;
> + int i, total_vqs;
> + char **names;
> +
> + /*
> + * We expect 1 RX virtqueue followed by 1 TX virtqueue, followd by
> + * possible control virtqueue, followed by RX/TX N-1 queue pairs used
> + * in multiqueue mode.
> + */
> + total_vqs = vi->max_queue_pairs * 2 +
> + virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ);
> +
> + /* Allocate space for find_vqs parameters */
> + vqs = kzalloc(total_vqs * sizeof(*vqs), GFP_KERNEL);
> + callbacks = kzalloc(total_vqs * sizeof(*callbacks), GFP_KERNEL);
> + if (!vqs || !callbacks)
> + goto err_mem;
> + names = kzalloc(total_vqs * sizeof(*names), GFP_KERNEL);
> + if (!names)
> + goto err_mem;
> +
> + /* Parameters for control virtqueue, if any */
> + if (vi->has_cvq) {
> + callbacks[2] = NULL;
> + names[2] = kasprintf(GFP_KERNEL, "control");
> + }
> +
> + /* Allocate/initialize parameters for send/receive virtqueues */
> + for (i = 0; i < vi->max_queue_pairs; i++) {
> + callbacks[rxq2vq(i)] = skb_recv_done;
> + callbacks[txq2vq(i)] = skb_xmit_done;
> + names[rxq2vq(i)] = kasprintf(GFP_KERNEL, "input.%d", i);
> + names[txq2vq(i)] = kasprintf(GFP_KERNEL, "output.%d", i);
> + }
> +
> + ret = vi->vdev->config->find_vqs(vi->vdev, total_vqs, vqs, callbacks,
> + (const char **)names);
> + if (ret)
> + goto err_names;
> +
> + if (vi->has_cvq) {
> + vi->cvq = vqs[2];
> if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VLAN))
> vi->dev->features |= NETIF_F_HW_VLAN_FILTER;
> }
> +
> + for (i = 0; i < vi->max_queue_pairs; i++) {
> + vi->rq[i].vq = vqs[rxq2vq(i)];
> + vi->sq[i].vq = vqs[txq2vq(i)];
> + }
> +
> + kfree(callbacks);
> + kfree(vqs);
> +
> + return 0;
> +
> +err_names:
> + for (i = 0; i < total_vqs * 2; i ++)
> + kfree(names[i]);
> + kfree(names);
> +
> +err_mem:
> + kfree(callbacks);
> + kfree(vqs);
> +
> + return ret;
> +}
> +
> +static int virtnet_alloc_queues(struct virtnet_info *vi)
> +{
> + int i;
> +
> + vi->sq = kzalloc(sizeof(vi->sq[0]) * vi->max_queue_pairs, GFP_KERNEL);
> + vi->rq = kzalloc(sizeof(vi->rq[0]) * vi->max_queue_pairs, GFP_KERNEL);
> + if (!vi->rq || !vi->sq)
> + goto err;
> +
> + /* setup initial receive and send queue parameters */
> + for (i = 0; i < vi->max_queue_pairs; i++) {
> + vi->rq[i].pages = NULL;
> + INIT_DELAYED_WORK(&vi->rq[i].refill, refill_work);
> + netif_napi_add(vi->dev, &vi->rq[i].napi, virtnet_poll,
> + napi_weight);
> +
> + sg_init_table(vi->rq[i].sg, ARRAY_SIZE(vi->rq[i].sg));
> + sg_init_table(vi->sq[i].sg, ARRAY_SIZE(vi->sq[i].sg));
> + }
> +
> +
> return 0;
> +
> +err:
> + virtnet_free_queues(vi);
> + return -ENOMEM;
> +}
> +
> +static int init_vqs(struct virtnet_info *vi)
> +{
> + int ret;
> +
> + /* Allocate send & receive queues */
> + ret = virtnet_alloc_queues(vi);
> + if (ret)
> + goto err;
> +
> + ret = virtnet_find_vqs(vi);
> + if (ret)
> + goto err_free;
> +
> + virtnet_set_affinity(vi, true);
> + return 0;
> +
> +err_free:
> + virtnet_free_queues(vi);
> +err:
> + return ret;
> }
>
> static int virtnet_probe(struct virtio_device *vdev)
> {
> - int err;
> + int i, err;
> struct net_device *dev;
> struct virtnet_info *vi;
> + u16 curr_queue_pairs;
Probably a good idea to rename this max_queue_pairs.
> +
> + /* Find if host supports multiqueue virtio_net device */
> + err = virtio_config_val(vdev, VIRTIO_NET_F_RFS,
> + offsetof(struct virtio_net_config,
> + max_virtqueue_pairs), &curr_queue_pairs);
> +
> + /* We need at least 2 queue's */
> + if (err)
> + curr_queue_pairs = 1;
Let's also validate against VIRTIO_NET_CTRL_RFS_VQ_PAIRS_MIN
and VIRTIO_NET_CTRL_RFS_VQ_PAIRS_MAX.
>
> /* Allocate ourselves a network device with room for our info */
> - dev = alloc_etherdev(sizeof(struct virtnet_info));
> + dev = alloc_etherdev_mq(sizeof(struct virtnet_info), curr_queue_pairs);
> if (!dev)
> return -ENOMEM;
>
> @@ -1126,22 +1374,17 @@ static int virtnet_probe(struct virtio_device *vdev)
>
> /* Set up our device-specific information */
> vi = netdev_priv(dev);
> - netif_napi_add(dev, &vi->rq.napi, virtnet_poll, napi_weight);
> vi->dev = dev;
> vi->vdev = vdev;
> vdev->priv = vi;
> - vi->rq.pages = NULL;
> vi->stats = alloc_percpu(struct virtnet_stats);
> err = -ENOMEM;
> if (vi->stats == NULL)
> goto free;
>
> - INIT_DELAYED_WORK(&vi->rq.refill, refill_work);
> mutex_init(&vi->config_lock);
> vi->config_enable = true;
> INIT_WORK(&vi->config_work, virtnet_config_changed_work);
> - sg_init_table(vi->rq.sg, ARRAY_SIZE(vi->rq.sg));
> - sg_init_table(vi->sq.sg, ARRAY_SIZE(vi->sq.sg));
>
> /* If we can receive ANY GSO packets, we must allocate large ones. */
> if (virtio_has_feature(vdev, VIRTIO_NET_F_GUEST_TSO4) ||
> @@ -1152,10 +1395,21 @@ static int virtnet_probe(struct virtio_device *vdev)
> if (virtio_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF))
> vi->mergeable_rx_bufs = true;
>
> + if (virtio_has_feature(vdev, VIRTIO_NET_F_CTRL_VQ))
> + vi->has_cvq = true;
> +
> + /* Use single tx/rx queue pair as default */
> + vi->curr_queue_pairs = 1;
> + vi->max_queue_pairs = curr_queue_pairs;
> +
> + /* Allocate/initialize the rx/tx queues, and invoke find_vqs */
> err = init_vqs(vi);
> if (err)
> goto free_stats;
>
> + netif_set_real_num_tx_queues(dev, 1);
> + netif_set_real_num_rx_queues(dev, 1);
> +
> err = register_netdev(dev);
> if (err) {
> pr_debug("virtio_net: registering device failed\n");
> @@ -1163,12 +1417,15 @@ static int virtnet_probe(struct virtio_device *vdev)
> }
>
> /* Last of all, set up some receive buffers. */
> - try_fill_recv(&vi->rq, GFP_KERNEL);
> -
> - /* If we didn't even get one input buffer, we're useless. */
> - if (vi->rq.num == 0) {
> - err = -ENOMEM;
> - goto unregister;
> + for (i = 0; i < vi->max_queue_pairs; i++) {
> + try_fill_recv(&vi->rq[i], GFP_KERNEL);
> +
> + /* If we didn't even get one input buffer, we're useless. */
> + if (vi->rq[i].num == 0) {
> + free_unused_bufs(vi);
> + err = -ENOMEM;
> + goto free_recv_bufs;
> + }
> }
>
> /* Assume link up if device can't report link status,
> @@ -1181,13 +1438,20 @@ static int virtnet_probe(struct virtio_device *vdev)
> netif_carrier_on(dev);
> }
>
> - pr_debug("virtnet: registered device %s\n", dev->name);
> + pr_debug("virtnet: registered device %s with %d RX and TX vq's\n",
> + dev->name, curr_queue_pairs);
> +
> return 0;
>
> -unregister:
> +free_recv_bufs:
> + free_receive_bufs(vi);
> unregister_netdev(dev);
> +
> free_vqs:
> - vdev->config->del_vqs(vdev);
> + for (i = 0; i <curr_queue_pairs; i++)
> + cancel_delayed_work_sync(&vi->rq[i].refill);
> + virtnet_del_vqs(vi);
> +
> free_stats:
> free_percpu(vi->stats);
> free:
> @@ -1195,28 +1459,6 @@ free:
> return err;
> }
>
> -static void free_unused_bufs(struct virtnet_info *vi)
> -{
> - void *buf;
> - while (1) {
> - buf = virtqueue_detach_unused_buf(vi->sq.vq);
> - if (!buf)
> - break;
> - dev_kfree_skb(buf);
> - }
> - while (1) {
> - buf = virtqueue_detach_unused_buf(vi->rq.vq);
> - if (!buf)
> - break;
> - if (vi->mergeable_rx_bufs || vi->big_packets)
> - give_pages(&vi->rq, buf);
> - else
> - dev_kfree_skb(buf);
> - --vi->rq.num;
> - }
> - BUG_ON(vi->rq.num != 0);
> -}
> -
> static void remove_vq_common(struct virtnet_info *vi)
> {
> vi->vdev->config->reset(vi->vdev);
> @@ -1224,10 +1466,9 @@ static void remove_vq_common(struct virtnet_info *vi)
> /* Free unused buffers in both send and recv, if any. */
> free_unused_bufs(vi);
>
> - vi->vdev->config->del_vqs(vi->vdev);
> + free_receive_bufs(vi);
>
> - while (vi->rq.pages)
> - __free_pages(get_a_page(&vi->rq, GFP_KERNEL), 0);
> + virtnet_del_vqs(vi);
> }
>
> static void __devexit virtnet_remove(struct virtio_device *vdev)
> @@ -1253,6 +1494,7 @@ static void __devexit virtnet_remove(struct virtio_device *vdev)
> static int virtnet_freeze(struct virtio_device *vdev)
> {
> struct virtnet_info *vi = vdev->priv;
> + int i;
>
> /* Prevent config work handler from accessing the device */
> mutex_lock(&vi->config_lock);
> @@ -1260,10 +1502,14 @@ static int virtnet_freeze(struct virtio_device *vdev)
> mutex_unlock(&vi->config_lock);
>
> netif_device_detach(vi->dev);
> - cancel_delayed_work_sync(&vi->rq.refill);
> + for (i = 0; i < vi->max_queue_pairs; i++)
> + cancel_delayed_work_sync(&vi->rq[i].refill);
>
> if (netif_running(vi->dev))
> - napi_disable(&vi->rq.napi);
> + for (i = 0; i < vi->max_queue_pairs; i++) {
> + napi_disable(&vi->rq[i].napi);
> + netif_napi_del(&vi->rq[i].napi);
> + }
>
> remove_vq_common(vi);
>
> @@ -1275,24 +1521,28 @@ static int virtnet_freeze(struct virtio_device *vdev)
> static int virtnet_restore(struct virtio_device *vdev)
> {
> struct virtnet_info *vi = vdev->priv;
> - int err;
> + int err, i;
>
> err = init_vqs(vi);
> if (err)
> return err;
>
> if (netif_running(vi->dev))
> - virtnet_napi_enable(&vi->rq);
> + for (i = 0; i < vi->max_queue_pairs; i++)
> + virtnet_napi_enable(&vi->rq[i]);
>
> netif_device_attach(vi->dev);
>
> - if (!try_fill_recv(&vi->rq, GFP_KERNEL))
> - schedule_delayed_work(&vi->rq.refill, 0);
> + for (i = 0; i < vi->max_queue_pairs; i++)
> + if (!try_fill_recv(&vi->rq[i], GFP_KERNEL))
> + schedule_delayed_work(&vi->rq[i].refill, 0);
>
> mutex_lock(&vi->config_lock);
> vi->config_enable = true;
> mutex_unlock(&vi->config_lock);
>
> + BUG_ON(virtnet_set_queues(vi));
> +
Won't this always fail when control vq is off?
> return 0;
> }
> #endif
> @@ -1310,7 +1560,7 @@ static unsigned int features[] = {
> VIRTIO_NET_F_GUEST_ECN, VIRTIO_NET_F_GUEST_UFO,
> VIRTIO_NET_F_MRG_RXBUF, VIRTIO_NET_F_STATUS, VIRTIO_NET_F_CTRL_VQ,
> VIRTIO_NET_F_CTRL_RX, VIRTIO_NET_F_CTRL_VLAN,
> - VIRTIO_NET_F_GUEST_ANNOUNCE,
> + VIRTIO_NET_F_GUEST_ANNOUNCE, VIRTIO_NET_F_RFS,
> };
>
> static struct virtio_driver virtio_net_driver = {
> @@ -1328,6 +1578,12 @@ static struct virtio_driver virtio_net_driver = {
> #endif
> };
>
> +static const struct ethtool_ops virtnet_ethtool_ops = {
> + .get_drvinfo = virtnet_get_drvinfo,
> + .get_link = ethtool_op_get_link,
> + .get_ringparam = virtnet_get_ringparam,
> +};
> +
> static int __init init(void)
> {
> return register_virtio_driver(&virtio_net_driver);
> diff --git a/include/uapi/linux/virtio_net.h b/include/uapi/linux/virtio_net.h
> index 2470f54..6056cec 100644
> --- a/include/uapi/linux/virtio_net.h
> +++ b/include/uapi/linux/virtio_net.h
> @@ -51,6 +51,7 @@
> #define VIRTIO_NET_F_CTRL_RX_EXTRA 20 /* Extra RX mode control support */
> #define VIRTIO_NET_F_GUEST_ANNOUNCE 21 /* Guest can announce device on the
> * network */
> +#define VIRTIO_NET_F_RFS 22 /* Device supports multiple TXQ/RXQ */
>
> #define VIRTIO_NET_S_LINK_UP 1 /* Link is up */
> #define VIRTIO_NET_S_ANNOUNCE 2 /* Announcement is needed */
> @@ -60,6 +61,8 @@ struct virtio_net_config {
> __u8 mac[6];
> /* See VIRTIO_NET_F_STATUS and VIRTIO_NET_S_* above */
> __u16 status;
> + /* Total number of RX/TX queues */
> + __u16 max_virtqueue_pairs;
> } __attribute__((packed));
>
> /* This is the first element of the scatter-gather list. If you don't
> @@ -166,4 +169,17 @@ struct virtio_net_ctrl_mac {
> #define VIRTIO_NET_CTRL_ANNOUNCE 3
> #define VIRTIO_NET_CTRL_ANNOUNCE_ACK 0
>
> +/*
> + * Control multiqueue
> + *
> + */
> +struct virtio_net_ctrl_rfs {
> + u16 virtqueue_pairs;
> +};
> +
> +#define VIRTIO_NET_CTRL_RFS 4
> + #define VIRTIO_NET_CTRL_RFS_VQ_PAIRS_SET 0
> + #define VIRTIO_NET_CTRL_RFS_VQ_PAIRS_MIN 1
> + #define VIRTIO_NET_CTRL_RFS_VQ_PAIRS_MAX 0x8000
> +
> #endif /* _LINUX_VIRTIO_NET_H */
> --
> 1.7.1
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 3/3] virtio-net: change the number of queues through ethtool
2012-11-27 10:16 ` [net-next rfc v7 3/3] virtio-net: change the number of queues through ethtool Jason Wang
@ 2012-12-02 16:09 ` Michael S. Tsirkin
2012-12-03 6:09 ` Jason Wang
0 siblings, 1 reply; 26+ messages in thread
From: Michael S. Tsirkin @ 2012-12-02 16:09 UTC (permalink / raw)
To: Jason Wang
Cc: rusty, krkumar2, virtualization, netdev, linux-kernel, kvm,
bhutchings, jwhan, shiyer
On Tue, Nov 27, 2012 at 06:16:00PM +0800, Jason Wang wrote:
> This patch implement the {set|get}_channels method of ethool to allow user to
> change the number of queues dymaically when the device is running. This would
> let the user to configure it on demand.
>
> Signed-off-by: Jason Wang <jasowang@redhat.com>
> ---
> drivers/net/virtio_net.c | 41 +++++++++++++++++++++++++++++++++++++++++
> 1 files changed, 41 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> index bcaa6e5..f08ec2a 100644
> --- a/drivers/net/virtio_net.c
> +++ b/drivers/net/virtio_net.c
> @@ -1578,10 +1578,51 @@ static struct virtio_driver virtio_net_driver = {
> #endif
> };
>
> +/* TODO: Eliminate OOO packets during switching */
> +static int virtnet_set_channels(struct net_device *dev,
> + struct ethtool_channels *channels)
> +{
> + struct virtnet_info *vi = netdev_priv(dev);
> + u16 queue_pairs = channels->combined_count;
> +
> + /* We don't support separate rx/tx channels.
> + * We don't allow setting 'other' channels.
> + */
> + if (channels->rx_count || channels->tx_count || channels->other_count)
> + return -EINVAL;
> +
> + /* Only two modes were support currently */
> + if (queue_pairs != vi->max_queue_pairs && queue_pairs != 1)
> + return -EINVAL;
> +
Why the limitation?
Also how does userspace discover what the legal values are?
> + vi->curr_queue_pairs = queue_pairs;
> + BUG_ON(virtnet_set_queues(vi));
> +
> + netif_set_real_num_tx_queues(dev, vi->curr_queue_pairs);
> + netif_set_real_num_rx_queues(dev, vi->curr_queue_pairs);
> +
> + return 0;
> +}
> +
> +static void virtnet_get_channels(struct net_device *dev,
> + struct ethtool_channels *channels)
> +{
> + struct virtnet_info *vi = netdev_priv(dev);
> +
> + channels->combined_count = vi->curr_queue_pairs;
> + channels->max_combined = vi->max_queue_pairs;
> + channels->max_other = 0;
> + channels->rx_count = 0;
> + channels->tx_count = 0;
> + channels->other_count = 0;
> +}
> +
> static const struct ethtool_ops virtnet_ethtool_ops = {
> .get_drvinfo = virtnet_get_drvinfo,
> .get_link = ethtool_op_get_link,
> .get_ringparam = virtnet_get_ringparam,
> + .set_channels = virtnet_set_channels,
> + .get_channels = virtnet_get_channels,
> };
>
> static int __init init(void)
> --
> 1.7.1
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 1/3] virtio-net: separate fields of sending/receiving queue from virtnet_info
2012-11-27 10:15 ` [net-next rfc v7 1/3] virtio-net: separate fields of sending/receiving queue from virtnet_info Jason Wang
@ 2012-12-03 1:55 ` Rusty Russell
[not found] ` <3524590.ZWGua7A8ne@jason-thinkpad-t430s>
0 siblings, 1 reply; 26+ messages in thread
From: Rusty Russell @ 2012-12-03 1:55 UTC (permalink / raw)
To: Jason Wang, mst, krkumar2, virtualization, netdev, linux-kernel
Cc: kvm, bhutchings, jwhan, shiyer, Jason Wang
Jason Wang <jasowang@redhat.com> writes:
> To support multiqueue transmitq/receiveq, the first step is to separate queue
> related structure from virtnet_info. This patch introduce send_queue and
> receive_queue structure and use the pointer to them as the parameter in
> functions handling sending/receiving.
OK, seems like a straightforward xform: a few nit-picks:
> +/* Internal representation of a receive virtqueue */
> +struct receive_queue {
> + /* Virtqueue associated with this receive_queue */
> + struct virtqueue *vq;
> +
> + struct napi_struct napi;
> +
> + /* Number of input buffers, and max we've ever had. */
> + unsigned int num, max;
Weird whitespace here.
> +
> + /* Work struct for refilling if we run low on memory. */
> + struct delayed_work refill;
I can't really see the justificaiton for a refill per queue. Just have
one work iterate all the queues if it happens, unless it happens often
(in which case, we need to look harder at this anyway).
> struct virtnet_info {
> struct virtio_device *vdev;
> - struct virtqueue *rvq, *svq, *cvq;
> + struct virtqueue *cvq;
> struct net_device *dev;
> struct napi_struct napi;
You leave napi here, and take it away in the next patch. I think it's
supposed to go away now.
Cheers,
Rusty.
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 2/3] virtio_net: multiqueue support
2012-11-27 10:15 ` [net-next rfc v7 2/3] virtio_net: multiqueue support Jason Wang
2012-12-02 16:06 ` Michael S. Tsirkin
@ 2012-12-03 2:04 ` Rusty Russell
2012-12-03 6:05 ` Jason Wang
2012-12-03 10:14 ` Michael S. Tsirkin
2 siblings, 1 reply; 26+ messages in thread
From: Rusty Russell @ 2012-12-03 2:04 UTC (permalink / raw)
To: Jason Wang, mst, krkumar2, virtualization, netdev, linux-kernel
Cc: kvm, bhutchings, jwhan, shiyer, Jason Wang
Jason Wang <jasowang@redhat.com> writes:
> +static const struct ethtool_ops virtnet_ethtool_ops;
> +
> +/*
> + * Converting between virtqueue no. and kernel tx/rx queue no.
> + * 0:rx0 1:tx0 2:cvq 3:rx1 4:tx1 ... 2N+1:rxN 2N+2:txN
> + */
> +static int vq2txq(struct virtqueue *vq)
> +{
> + int index = virtqueue_get_queue_index(vq);
> + return index == 1 ? 0 : (index - 2) / 2;
> +}
> +
> +static int txq2vq(int txq)
> +{
> + return txq ? 2 * txq + 2 : 1;
> +}
> +
> +static int vq2rxq(struct virtqueue *vq)
> +{
> + int index = virtqueue_get_queue_index(vq);
> + return index ? (index - 1) / 2 : 0;
> +}
> +
> +static int rxq2vq(int rxq)
> +{
> + return rxq ? 2 * rxq + 1 : 0;
> +}
> +
I thought MST changed the proposed spec to make the control queue always
the last one, so this logic becomes trivial.
> +static int virtnet_set_queues(struct virtnet_info *vi)
> +{
> + struct scatterlist sg;
> + struct virtio_net_ctrl_rfs s;
> + struct net_device *dev = vi->dev;
> +
> + s.virtqueue_pairs = vi->curr_queue_pairs;
> + sg_init_one(&sg, &s, sizeof(s));
> +
> + if (!vi->has_cvq)
> + return -EINVAL;
> +
> + if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RFS,
> + VIRTIO_NET_CTRL_RFS_VQ_PAIRS_SET, &sg, 1, 0)){
> + dev_warn(&dev->dev, "Fail to set the number of queue pairs to"
> + " %d\n", vi->curr_queue_pairs);
> + return -EINVAL;
> + }
Where do we check the VIRTIO_NET_F_RFS bit?
> static int virtnet_probe(struct virtio_device *vdev)
> {
> - int err;
> + int i, err;
> struct net_device *dev;
> struct virtnet_info *vi;
> + u16 curr_queue_pairs;
> +
> + /* Find if host supports multiqueue virtio_net device */
> + err = virtio_config_val(vdev, VIRTIO_NET_F_RFS,
> + offsetof(struct virtio_net_config,
> + max_virtqueue_pairs), &curr_queue_pairs);
> +
> + /* We need at least 2 queue's */
> + if (err)
> + curr_queue_pairs = 1;
Huh? Just call this queue_pairs. It's not curr_ at all...
> + if (virtio_has_feature(vdev, VIRTIO_NET_F_CTRL_VQ))
> + vi->has_cvq = true;
> +
> + /* Use single tx/rx queue pair as default */
> + vi->curr_queue_pairs = 1;
> + vi->max_queue_pairs = curr_queue_pairs;
See...
Cheers,
Rusty.
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 2/3] virtio_net: multiqueue support
2012-12-02 16:06 ` Michael S. Tsirkin
@ 2012-12-03 5:47 ` Jason Wang
0 siblings, 0 replies; 26+ messages in thread
From: Jason Wang @ 2012-12-03 5:47 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: rusty, krkumar2, virtualization, netdev, linux-kernel, kvm,
bhutchings, jwhan, shiyer
On Sunday, December 02, 2012 06:06:31 PM Michael S. Tsirkin wrote:
> On Tue, Nov 27, 2012 at 06:15:59PM +0800, Jason Wang wrote:
> > This addes multiqueue support to virtio_net driver. In multiple queue
> > modes, the driver expects the number of queue paris is equal to the
> > number of vcpus. To eliminate the contention bettwen vcpus and
> > virtqueues, per-cpu virtqueue pairs were implemented through:
> >
> > - select the txq based on the smp processor id.
> > - smp affinity hint were set to the vcpu that owns the queue pairs.
> >
> > Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>
> > Signed-off-by: Jason Wang <jasowang@redhat.com>
> > ---
> >
> > drivers/net/virtio_net.c | 454
> > ++++++++++++++++++++++++++++++--------- include/uapi/linux/virtio_net.h
> > | 16 ++
> > 2 files changed, 371 insertions(+), 99 deletions(-)
> >
> > diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> > index 7975133..bcaa6e5 100644
> > --- a/drivers/net/virtio_net.c
> > +++ b/drivers/net/virtio_net.c
> > @@ -84,17 +84,25 @@ struct virtnet_info {
> >
> > struct virtio_device *vdev;
> > struct virtqueue *cvq;
> > struct net_device *dev;
> >
> > - struct napi_struct napi;
> > - struct send_queue sq;
> > - struct receive_queue rq;
> > + struct send_queue *sq;
> > + struct receive_queue *rq;
> >
> > unsigned int status;
> >
> > + /* Max # of queue pairs supported by the device */
> > + u16 max_queue_pairs;
> > +
> > + /* # of queue pairs currently used by the driver */
> > + u16 curr_queue_pairs;
> > +
> >
> > /* I like... big packets and I cannot lie! */
> > bool big_packets;
> >
> > /* Host will merge rx buffers for big packets (shake it! shake it!) */
> > bool mergeable_rx_bufs;
> >
> > + /* Has control virtqueue */
> > + bool has_cvq;
> > +
> >
> > /* enable config space updates */
> > bool config_enable;
> >
> > @@ -126,6 +134,34 @@ struct padded_vnet_hdr {
> >
> > char padding[6];
> >
> > };
> >
> > +static const struct ethtool_ops virtnet_ethtool_ops;
> > +
> > +/*
> > + * Converting between virtqueue no. and kernel tx/rx queue no.
> > + * 0:rx0 1:tx0 2:cvq 3:rx1 4:tx1 ... 2N+1:rxN 2N+2:txN
> > + */
>
> Weird, this isn't what spec v5 says: it says
> 0:rx0 1:tx0 2: rx1 3: tx1 .... vcq
> We can change the spec to match but keeping all rx/tx
> together seems a bit prettier?
Oh, I miss the check of this part in v5. Have a thought about this, if we
change the location of cvq, we may break the support of legacy guest with only
single queue support. Consider we start a vm with 2 queue but boot a signle
queue legacy guest, it may think vq 2 is cvq which indeed is rx1.
>
> > +static int vq2txq(struct virtqueue *vq)
> > +{
> > + int index = virtqueue_get_queue_index(vq);
> > + return index == 1 ? 0 : (index - 2) / 2;
> > +}
> > +
> > +static int txq2vq(int txq)
> > +{
> > + return txq ? 2 * txq + 2 : 1;
> > +}
> > +
> > +static int vq2rxq(struct virtqueue *vq)
> > +{
> > + int index = virtqueue_get_queue_index(vq);
> > + return index ? (index - 1) / 2 : 0;
> > +}
> > +
> > +static int rxq2vq(int rxq)
> > +{
> > + return rxq ? 2 * rxq + 1 : 0;
> > +}
> > +
> >
> > static inline struct skb_vnet_hdr *skb_vnet_hdr(struct sk_buff *skb)
> > {
> >
> > return (struct skb_vnet_hdr *)skb->cb;
> >
> > @@ -166,7 +202,7 @@ static void skb_xmit_done(struct virtqueue *vq)
> >
> > virtqueue_disable_cb(vq);
> >
> > /* We were probably waiting for more output buffers. */
> >
> > - netif_wake_queue(vi->dev);
> > + netif_wake_subqueue(vi->dev, vq2txq(vq));
> >
> > }
> >
> > static void set_skb_frag(struct sk_buff *skb, struct page *page,
> >
> > @@ -503,7 +539,7 @@ static bool try_fill_recv(struct receive_queue *rq,
> > gfp_t gfp)>
> > static void skb_recv_done(struct virtqueue *rvq)
> > {
> >
> > struct virtnet_info *vi = rvq->vdev->priv;
> >
> > - struct receive_queue *rq = &vi->rq;
> > + struct receive_queue *rq = &vi->rq[vq2rxq(rvq)];
> >
> > /* Schedule NAPI, Suppress further interrupts if successful. */
> > if (napi_schedule_prep(&rq->napi)) {
> >
> > @@ -650,7 +686,8 @@ static int xmit_skb(struct send_queue *sq, struct
> > sk_buff *skb)>
> > static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device
> > *dev)
> > {
> >
> > struct virtnet_info *vi = netdev_priv(dev);
> >
> > - struct send_queue *sq = &vi->sq;
> > + int qnum = skb_get_queue_mapping(skb);
> > + struct send_queue *sq = &vi->sq[qnum];
> >
> > int capacity;
> >
> > /* Free up any pending old buffers before queueing new ones. */
> >
> > @@ -664,13 +701,14 @@ static netdev_tx_t start_xmit(struct sk_buff *skb,
> > struct net_device *dev)>
> > if (likely(capacity == -ENOMEM)) {
> >
> > if (net_ratelimit())
> >
> > dev_warn(&dev->dev,
> >
> > - "TX queue failure: out of memory\n");
> > + "TXQ (%d) failure: out of memory\n",
> > + qnum);
> >
> > } else {
> >
> > dev->stats.tx_fifo_errors++;
> > if (net_ratelimit())
> >
> > dev_warn(&dev->dev,
> >
> > - "Unexpected TX queue failure: %d\n",
> > - capacity);
> > + "Unexpected TXQ (%d) failure: %d\n",
> > + qnum, capacity);
> >
> > }
> > dev->stats.tx_dropped++;
> > kfree_skb(skb);
> >
> > @@ -685,12 +723,12 @@ static netdev_tx_t start_xmit(struct sk_buff *skb,
> > struct net_device *dev)>
> > /* Apparently nice girls don't return TX_BUSY; stop the queue
> >
> > * before it gets out of hand. Naturally, this wastes entries. */
> >
> > if (capacity < 2+MAX_SKB_FRAGS) {
> >
> > - netif_stop_queue(dev);
> > + netif_stop_subqueue(dev, qnum);
> >
> > if (unlikely(!virtqueue_enable_cb_delayed(sq->vq))) {
> >
> > /* More just got used, free them then recheck. */
> > capacity += free_old_xmit_skbs(sq);
> > if (capacity >= 2+MAX_SKB_FRAGS) {
> >
> > - netif_start_queue(dev);
> > + netif_start_subqueue(dev, qnum);
> >
> > virtqueue_disable_cb(sq->vq);
> >
> > }
> >
> > }
> >
> > @@ -758,23 +796,13 @@ static struct rtnl_link_stats64
> > *virtnet_stats(struct net_device *dev,>
> > static void virtnet_netpoll(struct net_device *dev)
> > {
> >
> > struct virtnet_info *vi = netdev_priv(dev);
> >
> > + int i;
> >
> > - napi_schedule(&vi->rq.napi);
> > + for (i = 0; i < vi->curr_queue_pairs; i++)
> > + napi_schedule(&vi->rq[i].napi);
> >
> > }
> > #endif
> >
> > -static int virtnet_open(struct net_device *dev)
> > -{
> > - struct virtnet_info *vi = netdev_priv(dev);
> > -
> > - /* Make sure we have some buffers: if oom use wq. */
> > - if (!try_fill_recv(&vi->rq, GFP_KERNEL))
> > - schedule_delayed_work(&vi->rq.refill, 0);
> > -
> > - virtnet_napi_enable(&vi->rq);
> > - return 0;
> > -}
> > -
> >
> > /*
> >
> > * Send command via the control virtqueue and check status. Commands
> > * supported by the hypervisor, as indicated by feature bits, should
> >
> > @@ -830,13 +858,53 @@ static void virtnet_ack_link_announce(struct
> > virtnet_info *vi)>
> > rtnl_unlock();
> >
> > }
> >
> > +static int virtnet_set_queues(struct virtnet_info *vi)
> > +{
> > + struct scatterlist sg;
> > + struct virtio_net_ctrl_rfs s;
> > + struct net_device *dev = vi->dev;
> > +
> > + s.virtqueue_pairs = vi->curr_queue_pairs;
> > + sg_init_one(&sg, &s, sizeof(s));
> > +
> > + if (!vi->has_cvq)
> > + return -EINVAL;
> > +
> > + if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RFS,
> > + VIRTIO_NET_CTRL_RFS_VQ_PAIRS_SET, &sg, 1, 0)){
> > + dev_warn(&dev->dev, "Fail to set the number of queue pairs to"
> > + " %d\n", vi->curr_queue_pairs);
> > + return -EINVAL;
> > + }
> > +
> > + return 0;
> > +}
> > +
> > +static int virtnet_open(struct net_device *dev)
> > +{
> > + struct virtnet_info *vi = netdev_priv(dev);
> > + int i;
> > +
> > + for (i = 0; i < vi->max_queue_pairs; i++) {
> > + /* Make sure we have some buffers: if oom use wq. */
> > + if (!try_fill_recv(&vi->rq[i], GFP_KERNEL))
> > + schedule_delayed_work(&vi->rq[i].refill, 0);
> > + virtnet_napi_enable(&vi->rq[i]);
> > + }
> > +
> > + return 0;
> > +}
> > +
> >
> > static int virtnet_close(struct net_device *dev)
> > {
> >
> > struct virtnet_info *vi = netdev_priv(dev);
> >
> > + int i;
> >
> > /* Make sure refill_work doesn't re-enable napi! */
> >
> > - cancel_delayed_work_sync(&vi->rq.refill);
> > - napi_disable(&vi->rq.napi);
> > + for (i = 0; i < vi->max_queue_pairs; i++) {
> > + cancel_delayed_work_sync(&vi->rq[i].refill);
> > + napi_disable(&vi->rq[i].napi);
> > + }
> >
> > return 0;
> >
> > }
> >
> > @@ -948,8 +1016,8 @@ static void virtnet_get_ringparam(struct net_device
> > *dev,>
> > {
> >
> > struct virtnet_info *vi = netdev_priv(dev);
> >
> > - ring->rx_max_pending = virtqueue_get_vring_size(vi->rq.vq);
> > - ring->tx_max_pending = virtqueue_get_vring_size(vi->sq.vq);
> > + ring->rx_max_pending = virtqueue_get_vring_size(vi->rq[0].vq);
> > + ring->tx_max_pending = virtqueue_get_vring_size(vi->sq[0].vq);
> >
> > ring->rx_pending = ring->rx_max_pending;
> > ring->tx_pending = ring->tx_max_pending;
> >
> > }
> >
> > @@ -967,12 +1035,6 @@ static void virtnet_get_drvinfo(struct net_device
> > *dev,>
> > }
> >
> > -static const struct ethtool_ops virtnet_ethtool_ops = {
> > - .get_drvinfo = virtnet_get_drvinfo,
> > - .get_link = ethtool_op_get_link,
> > - .get_ringparam = virtnet_get_ringparam,
> > -};
> > -
> >
> > #define MIN_MTU 68
> > #define MAX_MTU 65535
> >
> > @@ -984,6 +1046,20 @@ static int virtnet_change_mtu(struct net_device
> > *dev, int new_mtu)>
> > return 0;
> >
> > }
> >
> > +/* To avoid contending a lock hold by a vcpu who would exit to host,
> > select the + * txq based on the processor id.
> > + */
> > +static u16 virtnet_select_queue(struct net_device *dev, struct sk_buff
> > *skb) +{
> > + int txq = skb_rx_queue_recorded(skb) ? skb_get_rx_queue(skb) :
> > + smp_processor_id();
> > +
> > + while (unlikely(txq >= dev->real_num_tx_queues))
> > + txq -= dev->real_num_tx_queues;
> > +
> > + return txq;
> > +}
> > +
> >
> > static const struct net_device_ops virtnet_netdev = {
> >
> > .ndo_open = virtnet_open,
> > .ndo_stop = virtnet_close,
> >
> > @@ -995,6 +1071,7 @@ static const struct net_device_ops virtnet_netdev = {
> >
> > .ndo_get_stats64 = virtnet_stats,
> > .ndo_vlan_rx_add_vid = virtnet_vlan_rx_add_vid,
> > .ndo_vlan_rx_kill_vid = virtnet_vlan_rx_kill_vid,
> >
> > + .ndo_select_queue = virtnet_select_queue,
> >
> > #ifdef CONFIG_NET_POLL_CONTROLLER
> >
> > .ndo_poll_controller = virtnet_netpoll,
> >
> > #endif
> >
> > @@ -1030,10 +1107,10 @@ static void virtnet_config_changed_work(struct
> > work_struct *work)>
> > if (vi->status & VIRTIO_NET_S_LINK_UP) {
> >
> > netif_carrier_on(vi->dev);
> >
> > - netif_wake_queue(vi->dev);
> > + netif_tx_wake_all_queues(vi->dev);
> >
> > } else {
> >
> > netif_carrier_off(vi->dev);
> >
> > - netif_stop_queue(vi->dev);
> > + netif_tx_stop_all_queues(vi->dev);
> >
> > }
> >
> > done:
> > mutex_unlock(&vi->config_lock);
> >
> > @@ -1046,41 +1123,212 @@ static void virtnet_config_changed(struct
> > virtio_device *vdev)>
> > schedule_work(&vi->config_work);
> >
> > }
> >
> > -static int init_vqs(struct virtnet_info *vi)
> > +static void free_receive_bufs(struct virtnet_info *vi)
> > +{
> > + int i;
> > +
> > + for (i = 0; i < vi->max_queue_pairs; i++) {
> > + while (vi->rq[i].pages)
> > + __free_pages(get_a_page(&vi->rq[i], GFP_KERNEL), 0);
> > + }
> > +}
> > +
> > +/* Free memory allocated for send and receive queues */
> > +static void virtnet_free_queues(struct virtnet_info *vi)
> >
> > {
> >
> > - struct virtqueue *vqs[3];
> > - vq_callback_t *callbacks[] = { skb_recv_done, skb_xmit_done, NULL};
> > - const char *names[] = { "input", "output", "control" };
> > - int nvqs, err;
> > + kfree(vi->rq);
> > + vi->rq = NULL;
> > + kfree(vi->sq);
> > + vi->sq = NULL;
> > +}
> >
> > - /* We expect two virtqueues, receive then send,
> > - * and optionally control. */
> > - nvqs = virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ) ? 3 : 2;
> > +static void free_unused_bufs(struct virtnet_info *vi)
> > +{
> > + void *buf;
> > + int i;
> >
> > - err = vi->vdev->config->find_vqs(vi->vdev, nvqs, vqs, callbacks, names);
> > - if (err)
> > - return err;
> > + for (i = 0; i < vi->max_queue_pairs; i++) {
> > + struct virtqueue *vq = vi->sq[i].vq;
> > + while ((buf = virtqueue_detach_unused_buf(vq)) != NULL)
> > + dev_kfree_skb(buf);
> > + }
> >
> > - vi->rq.vq = vqs[0];
> > - vi->sq.vq = vqs[1];
> > + for (i = 0; i < vi->max_queue_pairs; i++) {
> > + struct virtqueue *vq = vi->rq[i].vq;
> >
> > - if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ)) {
> > - vi->cvq = vqs[2];
> > + while ((buf = virtqueue_detach_unused_buf(vq)) != NULL) {
> > + if (vi->mergeable_rx_bufs || vi->big_packets)
> > + give_pages(&vi->rq[i], buf);
> > + else
> > + dev_kfree_skb(buf);
> > + --vi->rq[i].num;
> > + }
> > + BUG_ON(vi->rq[i].num != 0);
> > + }
> > +}
> >
> > +static void virtnet_set_affinity(struct virtnet_info *vi, bool set)
> > +{
> > + int i;
> > +
> > + for (i = 0; i < vi->max_queue_pairs; i++) {
> > + int cpu = set ? i : -1;
> > + virtqueue_set_affinity(vi->rq[i].vq, cpu);
> > + virtqueue_set_affinity(vi->sq[i].vq, cpu);
> > + }
> > +}
> > +
> > +static void virtnet_del_vqs(struct virtnet_info *vi)
> > +{
> > + struct virtio_device *vdev = vi->vdev;
> > +
> > + virtnet_set_affinity(vi, false);
> > +
> > + vdev->config->del_vqs(vdev);
> > +
> > + virtnet_free_queues(vi);
> > +}
> > +
> > +static int virtnet_find_vqs(struct virtnet_info *vi)
> > +{
> > + vq_callback_t **callbacks;
> > + struct virtqueue **vqs;
> > + int ret = -ENOMEM;
> > + int i, total_vqs;
> > + char **names;
> > +
> > + /*
> > + * We expect 1 RX virtqueue followed by 1 TX virtqueue, followd by
> > + * possible control virtqueue, followed by RX/TX N-1 queue pairs used
> > + * in multiqueue mode.
> > + */
> > + total_vqs = vi->max_queue_pairs * 2 +
> > + virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ);
> > +
> > + /* Allocate space for find_vqs parameters */
> > + vqs = kzalloc(total_vqs * sizeof(*vqs), GFP_KERNEL);
> > + callbacks = kzalloc(total_vqs * sizeof(*callbacks), GFP_KERNEL);
> > + if (!vqs || !callbacks)
> > + goto err_mem;
> > + names = kzalloc(total_vqs * sizeof(*names), GFP_KERNEL);
> > + if (!names)
> > + goto err_mem;
> > +
> > + /* Parameters for control virtqueue, if any */
> > + if (vi->has_cvq) {
> > + callbacks[2] = NULL;
> > + names[2] = kasprintf(GFP_KERNEL, "control");
> > + }
> > +
> > + /* Allocate/initialize parameters for send/receive virtqueues */
> > + for (i = 0; i < vi->max_queue_pairs; i++) {
> > + callbacks[rxq2vq(i)] = skb_recv_done;
> > + callbacks[txq2vq(i)] = skb_xmit_done;
> > + names[rxq2vq(i)] = kasprintf(GFP_KERNEL, "input.%d", i);
> > + names[txq2vq(i)] = kasprintf(GFP_KERNEL, "output.%d", i);
> > + }
> > +
> > + ret = vi->vdev->config->find_vqs(vi->vdev, total_vqs, vqs, callbacks,
> > + (const char **)names);
> > + if (ret)
> > + goto err_names;
> > +
> > + if (vi->has_cvq) {
> > + vi->cvq = vqs[2];
> >
> > if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VLAN))
> >
> > vi->dev->features |= NETIF_F_HW_VLAN_FILTER;
> >
> > }
> >
> > +
> > + for (i = 0; i < vi->max_queue_pairs; i++) {
> > + vi->rq[i].vq = vqs[rxq2vq(i)];
> > + vi->sq[i].vq = vqs[txq2vq(i)];
> > + }
> > +
> > + kfree(callbacks);
> > + kfree(vqs);
> > +
> > + return 0;
> > +
> > +err_names:
> > + for (i = 0; i < total_vqs * 2; i ++)
> > + kfree(names[i]);
> > + kfree(names);
> > +
> > +err_mem:
> > + kfree(callbacks);
> > + kfree(vqs);
> > +
> > + return ret;
> > +}
> > +
> > +static int virtnet_alloc_queues(struct virtnet_info *vi)
> > +{
> > + int i;
> > +
> > + vi->sq = kzalloc(sizeof(vi->sq[0]) * vi->max_queue_pairs, GFP_KERNEL);
> > + vi->rq = kzalloc(sizeof(vi->rq[0]) * vi->max_queue_pairs, GFP_KERNEL);
> > + if (!vi->rq || !vi->sq)
> > + goto err;
> > +
> > + /* setup initial receive and send queue parameters */
> > + for (i = 0; i < vi->max_queue_pairs; i++) {
> > + vi->rq[i].pages = NULL;
> > + INIT_DELAYED_WORK(&vi->rq[i].refill, refill_work);
> > + netif_napi_add(vi->dev, &vi->rq[i].napi, virtnet_poll,
> > + napi_weight);
> > +
> > + sg_init_table(vi->rq[i].sg, ARRAY_SIZE(vi->rq[i].sg));
> > + sg_init_table(vi->sq[i].sg, ARRAY_SIZE(vi->sq[i].sg));
> > + }
> > +
> > +
> >
> > return 0;
> >
> > +
> > +err:
> > + virtnet_free_queues(vi);
> > + return -ENOMEM;
> > +}
> > +
> > +static int init_vqs(struct virtnet_info *vi)
> > +{
> > + int ret;
> > +
> > + /* Allocate send & receive queues */
> > + ret = virtnet_alloc_queues(vi);
> > + if (ret)
> > + goto err;
> > +
> > + ret = virtnet_find_vqs(vi);
> > + if (ret)
> > + goto err_free;
> > +
> > + virtnet_set_affinity(vi, true);
> > + return 0;
> > +
> > +err_free:
> > + virtnet_free_queues(vi);
> > +err:
> > + return ret;
> >
> > }
> >
> > static int virtnet_probe(struct virtio_device *vdev)
> > {
> >
> > - int err;
> > + int i, err;
> >
> > struct net_device *dev;
> > struct virtnet_info *vi;
> >
> > + u16 curr_queue_pairs;
>
> Probably a good idea to rename this max_queue_pairs.
Sure.
>
> > +
> > + /* Find if host supports multiqueue virtio_net device */
> > + err = virtio_config_val(vdev, VIRTIO_NET_F_RFS,
> > + offsetof(struct virtio_net_config,
> > + max_virtqueue_pairs), &curr_queue_pairs);
> > +
> > + /* We need at least 2 queue's */
> > + if (err)
> > + curr_queue_pairs = 1;
>
> Let's also validate against VIRTIO_NET_CTRL_RFS_VQ_PAIRS_MIN
> and VIRTIO_NET_CTRL_RFS_VQ_PAIRS_MAX.
Ok.
>
> > /* Allocate ourselves a network device with room for our info */
> >
> > - dev = alloc_etherdev(sizeof(struct virtnet_info));
> > + dev = alloc_etherdev_mq(sizeof(struct virtnet_info), curr_queue_pairs);
> >
> > if (!dev)
> >
> > return -ENOMEM;
> >
> > @@ -1126,22 +1374,17 @@ static int virtnet_probe(struct virtio_device
> > *vdev)>
> > /* Set up our device-specific information */
> > vi = netdev_priv(dev);
> >
> > - netif_napi_add(dev, &vi->rq.napi, virtnet_poll, napi_weight);
> >
> > vi->dev = dev;
> > vi->vdev = vdev;
> > vdev->priv = vi;
> >
> > - vi->rq.pages = NULL;
> >
> > vi->stats = alloc_percpu(struct virtnet_stats);
> > err = -ENOMEM;
> > if (vi->stats == NULL)
> >
> > goto free;
> >
> > - INIT_DELAYED_WORK(&vi->rq.refill, refill_work);
> >
> > mutex_init(&vi->config_lock);
> > vi->config_enable = true;
> > INIT_WORK(&vi->config_work, virtnet_config_changed_work);
> >
> > - sg_init_table(vi->rq.sg, ARRAY_SIZE(vi->rq.sg));
> > - sg_init_table(vi->sq.sg, ARRAY_SIZE(vi->sq.sg));
> >
> > /* If we can receive ANY GSO packets, we must allocate large ones. */
> > if (virtio_has_feature(vdev, VIRTIO_NET_F_GUEST_TSO4) ||
> >
> > @@ -1152,10 +1395,21 @@ static int virtnet_probe(struct virtio_device
> > *vdev)>
> > if (virtio_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF))
> >
> > vi->mergeable_rx_bufs = true;
> >
> > + if (virtio_has_feature(vdev, VIRTIO_NET_F_CTRL_VQ))
> > + vi->has_cvq = true;
> > +
> > + /* Use single tx/rx queue pair as default */
> > + vi->curr_queue_pairs = 1;
> > + vi->max_queue_pairs = curr_queue_pairs;
> > +
> > + /* Allocate/initialize the rx/tx queues, and invoke find_vqs */
> >
> > err = init_vqs(vi);
> > if (err)
> >
> > goto free_stats;
> >
> > + netif_set_real_num_tx_queues(dev, 1);
> > + netif_set_real_num_rx_queues(dev, 1);
> > +
> >
> > err = register_netdev(dev);
> > if (err) {
> >
> > pr_debug("virtio_net: registering device failed\n");
> >
> > @@ -1163,12 +1417,15 @@ static int virtnet_probe(struct virtio_device
> > *vdev)>
> > }
> >
> > /* Last of all, set up some receive buffers. */
> >
> > - try_fill_recv(&vi->rq, GFP_KERNEL);
> > -
> > - /* If we didn't even get one input buffer, we're useless. */
> > - if (vi->rq.num == 0) {
> > - err = -ENOMEM;
> > - goto unregister;
> > + for (i = 0; i < vi->max_queue_pairs; i++) {
> > + try_fill_recv(&vi->rq[i], GFP_KERNEL);
> > +
> > + /* If we didn't even get one input buffer, we're useless. */
> > + if (vi->rq[i].num == 0) {
> > + free_unused_bufs(vi);
> > + err = -ENOMEM;
> > + goto free_recv_bufs;
> > + }
> >
> > }
> >
> > /* Assume link up if device can't report link status,
> >
> > @@ -1181,13 +1438,20 @@ static int virtnet_probe(struct virtio_device
> > *vdev)>
> > netif_carrier_on(dev);
> >
> > }
> >
> > - pr_debug("virtnet: registered device %s\n", dev->name);
> > + pr_debug("virtnet: registered device %s with %d RX and TX vq's\n",
> > + dev->name, curr_queue_pairs);
> > +
> >
> > return 0;
> >
> > -unregister:
> > +free_recv_bufs:
> > + free_receive_bufs(vi);
> >
> > unregister_netdev(dev);
> >
> > +
> >
> > free_vqs:
> > - vdev->config->del_vqs(vdev);
> > + for (i = 0; i <curr_queue_pairs; i++)
> > + cancel_delayed_work_sync(&vi->rq[i].refill);
> > + virtnet_del_vqs(vi);
> > +
> >
> > free_stats:
> > free_percpu(vi->stats);
> >
> > free:
> > @@ -1195,28 +1459,6 @@ free:
> > return err;
> >
> > }
> >
> > -static void free_unused_bufs(struct virtnet_info *vi)
> > -{
> > - void *buf;
> > - while (1) {
> > - buf = virtqueue_detach_unused_buf(vi->sq.vq);
> > - if (!buf)
> > - break;
> > - dev_kfree_skb(buf);
> > - }
> > - while (1) {
> > - buf = virtqueue_detach_unused_buf(vi->rq.vq);
> > - if (!buf)
> > - break;
> > - if (vi->mergeable_rx_bufs || vi->big_packets)
> > - give_pages(&vi->rq, buf);
> > - else
> > - dev_kfree_skb(buf);
> > - --vi->rq.num;
> > - }
> > - BUG_ON(vi->rq.num != 0);
> > -}
> > -
> >
> > static void remove_vq_common(struct virtnet_info *vi)
> > {
> >
> > vi->vdev->config->reset(vi->vdev);
> >
> > @@ -1224,10 +1466,9 @@ static void remove_vq_common(struct virtnet_info
> > *vi)>
> > /* Free unused buffers in both send and recv, if any. */
> > free_unused_bufs(vi);
> >
> > - vi->vdev->config->del_vqs(vi->vdev);
> > + free_receive_bufs(vi);
> >
> > - while (vi->rq.pages)
> > - __free_pages(get_a_page(&vi->rq, GFP_KERNEL), 0);
> > + virtnet_del_vqs(vi);
> >
> > }
> >
> > static void __devexit virtnet_remove(struct virtio_device *vdev)
> >
> > @@ -1253,6 +1494,7 @@ static void __devexit virtnet_remove(struct
> > virtio_device *vdev)>
> > static int virtnet_freeze(struct virtio_device *vdev)
> > {
> >
> > struct virtnet_info *vi = vdev->priv;
> >
> > + int i;
> >
> > /* Prevent config work handler from accessing the device */
> > mutex_lock(&vi->config_lock);
> >
> > @@ -1260,10 +1502,14 @@ static int virtnet_freeze(struct virtio_device
> > *vdev)>
> > mutex_unlock(&vi->config_lock);
> >
> > netif_device_detach(vi->dev);
> >
> > - cancel_delayed_work_sync(&vi->rq.refill);
> > + for (i = 0; i < vi->max_queue_pairs; i++)
> > + cancel_delayed_work_sync(&vi->rq[i].refill);
> >
> > if (netif_running(vi->dev))
> >
> > - napi_disable(&vi->rq.napi);
> > + for (i = 0; i < vi->max_queue_pairs; i++) {
> > + napi_disable(&vi->rq[i].napi);
> > + netif_napi_del(&vi->rq[i].napi);
> > + }
> >
> > remove_vq_common(vi);
> >
> > @@ -1275,24 +1521,28 @@ static int virtnet_freeze(struct virtio_device
> > *vdev)>
> > static int virtnet_restore(struct virtio_device *vdev)
> > {
> >
> > struct virtnet_info *vi = vdev->priv;
> >
> > - int err;
> > + int err, i;
> >
> > err = init_vqs(vi);
> > if (err)
> >
> > return err;
> >
> > if (netif_running(vi->dev))
> >
> > - virtnet_napi_enable(&vi->rq);
> > + for (i = 0; i < vi->max_queue_pairs; i++)
> > + virtnet_napi_enable(&vi->rq[i]);
> >
> > netif_device_attach(vi->dev);
> >
> > - if (!try_fill_recv(&vi->rq, GFP_KERNEL))
> > - schedule_delayed_work(&vi->rq.refill, 0);
> > + for (i = 0; i < vi->max_queue_pairs; i++)
> > + if (!try_fill_recv(&vi->rq[i], GFP_KERNEL))
> > + schedule_delayed_work(&vi->rq[i].refill, 0);
> >
> > mutex_lock(&vi->config_lock);
> > vi->config_enable = true;
> > mutex_unlock(&vi->config_lock);
> >
> > + BUG_ON(virtnet_set_queues(vi));
> > +
>
> Won't this always fail when control vq is off?
Yes, will add a check of VIRTIO_NET_F_RFS before calling virtnet_set_queues().
>
> > return 0;
> >
> > }
> > #endif
> >
> > @@ -1310,7 +1560,7 @@ static unsigned int features[] = {
> >
> > VIRTIO_NET_F_GUEST_ECN, VIRTIO_NET_F_GUEST_UFO,
> > VIRTIO_NET_F_MRG_RXBUF, VIRTIO_NET_F_STATUS, VIRTIO_NET_F_CTRL_VQ,
> > VIRTIO_NET_F_CTRL_RX, VIRTIO_NET_F_CTRL_VLAN,
> >
> > - VIRTIO_NET_F_GUEST_ANNOUNCE,
> > + VIRTIO_NET_F_GUEST_ANNOUNCE, VIRTIO_NET_F_RFS,
> >
> > };
> >
> > static struct virtio_driver virtio_net_driver = {
> >
> > @@ -1328,6 +1578,12 @@ static struct virtio_driver virtio_net_driver = {
> >
> > #endif
> > };
> >
> > +static const struct ethtool_ops virtnet_ethtool_ops = {
> > + .get_drvinfo = virtnet_get_drvinfo,
> > + .get_link = ethtool_op_get_link,
> > + .get_ringparam = virtnet_get_ringparam,
> > +};
> > +
> >
> > static int __init init(void)
> > {
> >
> > return register_virtio_driver(&virtio_net_driver);
> >
> > diff --git a/include/uapi/linux/virtio_net.h
> > b/include/uapi/linux/virtio_net.h index 2470f54..6056cec 100644
> > --- a/include/uapi/linux/virtio_net.h
> > +++ b/include/uapi/linux/virtio_net.h
> > @@ -51,6 +51,7 @@
> >
> > #define VIRTIO_NET_F_CTRL_RX_EXTRA 20 /* Extra RX mode control support
*/
> > #define VIRTIO_NET_F_GUEST_ANNOUNCE 21 /* Guest can announce device on
> > the
> >
> > * network */
> >
> > +#define VIRTIO_NET_F_RFS 22 /* Device supports multiple TXQ/RXQ */
> >
> > #define VIRTIO_NET_S_LINK_UP 1 /* Link is up */
> > #define VIRTIO_NET_S_ANNOUNCE 2 /* Announcement is needed */
> >
> > @@ -60,6 +61,8 @@ struct virtio_net_config {
> >
> > __u8 mac[6];
> > /* See VIRTIO_NET_F_STATUS and VIRTIO_NET_S_* above */
> > __u16 status;
> >
> > + /* Total number of RX/TX queues */
> > + __u16 max_virtqueue_pairs;
> >
> > } __attribute__((packed));
> >
> > /* This is the first element of the scatter-gather list. If you don't
> >
> > @@ -166,4 +169,17 @@ struct virtio_net_ctrl_mac {
> >
> > #define VIRTIO_NET_CTRL_ANNOUNCE 3
> >
> > #define VIRTIO_NET_CTRL_ANNOUNCE_ACK 0
> >
> > +/*
> > + * Control multiqueue
> > + *
> > + */
> > +struct virtio_net_ctrl_rfs {
> > + u16 virtqueue_pairs;
> > +};
> > +
> > +#define VIRTIO_NET_CTRL_RFS 4
> > + #define VIRTIO_NET_CTRL_RFS_VQ_PAIRS_SET 0
> > + #define VIRTIO_NET_CTRL_RFS_VQ_PAIRS_MIN 1
> > + #define VIRTIO_NET_CTRL_RFS_VQ_PAIRS_MAX 0x8000
> > +
> >
> > #endif /* _LINUX_VIRTIO_NET_H */
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 2/3] virtio_net: multiqueue support
2012-12-03 2:04 ` Rusty Russell
@ 2012-12-03 6:05 ` Jason Wang
2012-12-03 9:47 ` Michael S. Tsirkin
0 siblings, 1 reply; 26+ messages in thread
From: Jason Wang @ 2012-12-03 6:05 UTC (permalink / raw)
To: Rusty Russell
Cc: mst, krkumar2, virtualization, netdev, linux-kernel, kvm,
bhutchings, jwhan, shiyer
On Monday, December 03, 2012 12:34:08 PM Rusty Russell wrote:
> Jason Wang <jasowang@redhat.com> writes:
> > +static const struct ethtool_ops virtnet_ethtool_ops;
> > +
> > +/*
> > + * Converting between virtqueue no. and kernel tx/rx queue no.
> > + * 0:rx0 1:tx0 2:cvq 3:rx1 4:tx1 ... 2N+1:rxN 2N+2:txN
> > + */
> > +static int vq2txq(struct virtqueue *vq)
> > +{
> > + int index = virtqueue_get_queue_index(vq);
> > + return index == 1 ? 0 : (index - 2) / 2;
> > +}
> > +
> > +static int txq2vq(int txq)
> > +{
> > + return txq ? 2 * txq + 2 : 1;
> > +}
> > +
> > +static int vq2rxq(struct virtqueue *vq)
> > +{
> > + int index = virtqueue_get_queue_index(vq);
> > + return index ? (index - 1) / 2 : 0;
> > +}
> > +
> > +static int rxq2vq(int rxq)
> > +{
> > + return rxq ? 2 * rxq + 1 : 0;
> > +}
> > +
>
> I thought MST changed the proposed spec to make the control queue always
> the last one, so this logic becomes trivial.
But it may break the support of legacy guest. If we boot a legacy single queue
guest on a 2 queue virtio-net device. It may think vq 2 is cvq which is indeed
rx1.
>
> > +static int virtnet_set_queues(struct virtnet_info *vi)
> > +{
> > + struct scatterlist sg;
> > + struct virtio_net_ctrl_rfs s;
> > + struct net_device *dev = vi->dev;
> > +
> > + s.virtqueue_pairs = vi->curr_queue_pairs;
> > + sg_init_one(&sg, &s, sizeof(s));
> > +
> > + if (!vi->has_cvq)
> > + return -EINVAL;
> > +
> > + if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RFS,
> > + VIRTIO_NET_CTRL_RFS_VQ_PAIRS_SET, &sg, 1, 0)){
> > + dev_warn(&dev->dev, "Fail to set the number of queue pairs to"
> > + " %d\n", vi->curr_queue_pairs);
> > + return -EINVAL;
> > + }
>
> Where do we check the VIRTIO_NET_F_RFS bit?
Yes, we need this check. Will let the caller does the check and add a comment
and check in the caller.
>
> > static int virtnet_probe(struct virtio_device *vdev)
> > {
> >
> > - int err;
> > + int i, err;
> >
> > struct net_device *dev;
> > struct virtnet_info *vi;
> >
> > + u16 curr_queue_pairs;
> > +
> > + /* Find if host supports multiqueue virtio_net device */
> > + err = virtio_config_val(vdev, VIRTIO_NET_F_RFS,
> > + offsetof(struct virtio_net_config,
> > + max_virtqueue_pairs), &curr_queue_pairs);
> > +
> > + /* We need at least 2 queue's */
> > + if (err)
> > + curr_queue_pairs = 1;
>
> Huh? Just call this queue_pairs. It's not curr_ at all...
>
> > + if (virtio_has_feature(vdev, VIRTIO_NET_F_CTRL_VQ))
> > + vi->has_cvq = true;
> > +
> > + /* Use single tx/rx queue pair as default */
> > + vi->curr_queue_pairs = 1;
> > + vi->max_queue_pairs = curr_queue_pairs;
>
> See...
Right, will use max_queue_pairs then.
Thanks
>
> Cheers,
> Rusty.
> --
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 3/3] virtio-net: change the number of queues through ethtool
2012-12-02 16:09 ` Michael S. Tsirkin
@ 2012-12-03 6:09 ` Jason Wang
2012-12-03 11:25 ` Michael S. Tsirkin
0 siblings, 1 reply; 26+ messages in thread
From: Jason Wang @ 2012-12-03 6:09 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: rusty, krkumar2, virtualization, netdev, linux-kernel, kvm,
bhutchings, jwhan, shiyer
On Sunday, December 02, 2012 06:09:06 PM Michael S. Tsirkin wrote:
> On Tue, Nov 27, 2012 at 06:16:00PM +0800, Jason Wang wrote:
> > This patch implement the {set|get}_channels method of ethool to allow user
> > to change the number of queues dymaically when the device is running.
> > This would let the user to configure it on demand.
> >
> > Signed-off-by: Jason Wang <jasowang@redhat.com>
> > ---
> >
> > drivers/net/virtio_net.c | 41 +++++++++++++++++++++++++++++++++++++++++
> > 1 files changed, 41 insertions(+), 0 deletions(-)
> >
> > diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> > index bcaa6e5..f08ec2a 100644
> > --- a/drivers/net/virtio_net.c
> > +++ b/drivers/net/virtio_net.c
> > @@ -1578,10 +1578,51 @@ static struct virtio_driver virtio_net_driver = {
> >
> > #endif
> > };
> >
> > +/* TODO: Eliminate OOO packets during switching */
> > +static int virtnet_set_channels(struct net_device *dev,
> > + struct ethtool_channels *channels)
> > +{
> > + struct virtnet_info *vi = netdev_priv(dev);
> > + u16 queue_pairs = channels->combined_count;
> > +
> > + /* We don't support separate rx/tx channels.
> > + * We don't allow setting 'other' channels.
> > + */
> > + if (channels->rx_count || channels->tx_count || channels->other_count)
> > + return -EINVAL;
> > +
> > + /* Only two modes were support currently */
> > + if (queue_pairs != vi->max_queue_pairs && queue_pairs != 1)
> > + return -EINVAL;
> > +
>
> Why the limitation?
Not sure the value bettwen 1 and max_queue_pairs is useful. But anyway, I can
remove this limitation.
> Also how does userspace discover what the legal values are?
Userspace only check whether the value is greater than max_queue_pairs.
>
> > + vi->curr_queue_pairs = queue_pairs;
> > + BUG_ON(virtnet_set_queues(vi));
> > +
> > + netif_set_real_num_tx_queues(dev, vi->curr_queue_pairs);
> > + netif_set_real_num_rx_queues(dev, vi->curr_queue_pairs);
> > +
> > + return 0;
> > +}
> > +
> > +static void virtnet_get_channels(struct net_device *dev,
> > + struct ethtool_channels *channels)
> > +{
> > + struct virtnet_info *vi = netdev_priv(dev);
> > +
> > + channels->combined_count = vi->curr_queue_pairs;
> > + channels->max_combined = vi->max_queue_pairs;
> > + channels->max_other = 0;
> > + channels->rx_count = 0;
> > + channels->tx_count = 0;
> > + channels->other_count = 0;
> > +}
> > +
> >
> > static const struct ethtool_ops virtnet_ethtool_ops = {
> >
> > .get_drvinfo = virtnet_get_drvinfo,
> > .get_link = ethtool_op_get_link,
> > .get_ringparam = virtnet_get_ringparam,
> >
> > + .set_channels = virtnet_set_channels,
> > + .get_channels = virtnet_get_channels,
> >
> > };
> >
> > static int __init init(void)
>
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 2/3] virtio_net: multiqueue support
2012-12-03 6:05 ` Jason Wang
@ 2012-12-03 9:47 ` Michael S. Tsirkin
2012-12-03 10:01 ` Jason Wang
0 siblings, 1 reply; 26+ messages in thread
From: Michael S. Tsirkin @ 2012-12-03 9:47 UTC (permalink / raw)
To: Jason Wang
Cc: Rusty Russell, krkumar2, virtualization, netdev, linux-kernel,
kvm, bhutchings, jwhan, shiyer
On Mon, Dec 03, 2012 at 02:05:27PM +0800, Jason Wang wrote:
> On Monday, December 03, 2012 12:34:08 PM Rusty Russell wrote:
> > Jason Wang <jasowang@redhat.com> writes:
> > > +static const struct ethtool_ops virtnet_ethtool_ops;
> > > +
> > > +/*
> > > + * Converting between virtqueue no. and kernel tx/rx queue no.
> > > + * 0:rx0 1:tx0 2:cvq 3:rx1 4:tx1 ... 2N+1:rxN 2N+2:txN
> > > + */
> > > +static int vq2txq(struct virtqueue *vq)
> > > +{
> > > + int index = virtqueue_get_queue_index(vq);
> > > + return index == 1 ? 0 : (index - 2) / 2;
> > > +}
> > > +
> > > +static int txq2vq(int txq)
> > > +{
> > > + return txq ? 2 * txq + 2 : 1;
> > > +}
> > > +
> > > +static int vq2rxq(struct virtqueue *vq)
> > > +{
> > > + int index = virtqueue_get_queue_index(vq);
> > > + return index ? (index - 1) / 2 : 0;
> > > +}
> > > +
> > > +static int rxq2vq(int rxq)
> > > +{
> > > + return rxq ? 2 * rxq + 1 : 0;
> > > +}
> > > +
> >
> > I thought MST changed the proposed spec to make the control queue always
> > the last one, so this logic becomes trivial.
>
> But it may break the support of legacy guest. If we boot a legacy single queue
> guest on a 2 queue virtio-net device. It may think vq 2 is cvq which is indeed
> rx1.
Legacy guyest support should be handled by host using feature
bits in the usual way: host should detect legacy guest
by checking the VIRTIO_NET_F_RFS feature.
If VIRTIO_NET_F_RFS is acked, cvq is vq max_virtqueue_pairs * 2.
If it's not acked, cvq is vq 2.
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 2/3] virtio_net: multiqueue support
2012-12-03 9:47 ` Michael S. Tsirkin
@ 2012-12-03 10:01 ` Jason Wang
2012-12-03 11:11 ` Michael S. Tsirkin
0 siblings, 1 reply; 26+ messages in thread
From: Jason Wang @ 2012-12-03 10:01 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: Rusty Russell, krkumar2, virtualization, netdev, linux-kernel,
kvm, bhutchings, jwhan, shiyer
On 12/03/2012 05:47 PM, Michael S. Tsirkin wrote:
> On Mon, Dec 03, 2012 at 02:05:27PM +0800, Jason Wang wrote:
>> On Monday, December 03, 2012 12:34:08 PM Rusty Russell wrote:
>>> Jason Wang <jasowang@redhat.com> writes:
>>>> +static const struct ethtool_ops virtnet_ethtool_ops;
>>>> +
>>>> +/*
>>>> + * Converting between virtqueue no. and kernel tx/rx queue no.
>>>> + * 0:rx0 1:tx0 2:cvq 3:rx1 4:tx1 ... 2N+1:rxN 2N+2:txN
>>>> + */
>>>> +static int vq2txq(struct virtqueue *vq)
>>>> +{
>>>> + int index = virtqueue_get_queue_index(vq);
>>>> + return index == 1 ? 0 : (index - 2) / 2;
>>>> +}
>>>> +
>>>> +static int txq2vq(int txq)
>>>> +{
>>>> + return txq ? 2 * txq + 2 : 1;
>>>> +}
>>>> +
>>>> +static int vq2rxq(struct virtqueue *vq)
>>>> +{
>>>> + int index = virtqueue_get_queue_index(vq);
>>>> + return index ? (index - 1) / 2 : 0;
>>>> +}
>>>> +
>>>> +static int rxq2vq(int rxq)
>>>> +{
>>>> + return rxq ? 2 * rxq + 1 : 0;
>>>> +}
>>>> +
>>> I thought MST changed the proposed spec to make the control queue always
>>> the last one, so this logic becomes trivial.
>> But it may break the support of legacy guest. If we boot a legacy single queue
>> guest on a 2 queue virtio-net device. It may think vq 2 is cvq which is indeed
>> rx1.
> Legacy guyest support should be handled by host using feature
> bits in the usual way: host should detect legacy guest
> by checking the VIRTIO_NET_F_RFS feature.
>
> If VIRTIO_NET_F_RFS is acked, cvq is vq max_virtqueue_pairs * 2.
> If it's not acked, cvq is vq 2.
>
We could, but we didn't gain much from this. Furthermore, we need also do the dynamic creation/destroying of virtqueues during feature negotiation which seems not supported in qemu now.
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 2/3] virtio_net: multiqueue support
2012-11-27 10:15 ` [net-next rfc v7 2/3] virtio_net: multiqueue support Jason Wang
2012-12-02 16:06 ` Michael S. Tsirkin
2012-12-03 2:04 ` Rusty Russell
@ 2012-12-03 10:14 ` Michael S. Tsirkin
2012-12-03 10:30 ` Jason Wang
2 siblings, 1 reply; 26+ messages in thread
From: Michael S. Tsirkin @ 2012-12-03 10:14 UTC (permalink / raw)
To: Jason Wang
Cc: rusty, krkumar2, virtualization, netdev, linux-kernel, kvm,
bhutchings, jwhan, shiyer
On Tue, Nov 27, 2012 at 06:15:59PM +0800, Jason Wang wrote:
> - if (!try_fill_recv(&vi->rq, GFP_KERNEL))
> - schedule_delayed_work(&vi->rq.refill, 0);
> + for (i = 0; i < vi->max_queue_pairs; i++)
> + if (!try_fill_recv(&vi->rq[i], GFP_KERNEL))
> + schedule_delayed_work(&vi->rq[i].refill, 0);
>
> mutex_lock(&vi->config_lock);
> vi->config_enable = true;
> mutex_unlock(&vi->config_lock);
>
> + BUG_ON(virtnet_set_queues(vi));
> +
> return 0;
> }
> #endif
Also crashing on device nack of command is also not nice.
In this case it seems we can just switch to
single-queue mode which should always be safe.
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 2/3] virtio_net: multiqueue support
2012-12-03 10:14 ` Michael S. Tsirkin
@ 2012-12-03 10:30 ` Jason Wang
2012-12-03 11:00 ` Michael S. Tsirkin
2012-12-04 7:35 ` Michael S. Tsirkin
0 siblings, 2 replies; 26+ messages in thread
From: Jason Wang @ 2012-12-03 10:30 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: rusty, krkumar2, virtualization, netdev, linux-kernel, kvm,
bhutchings, jwhan, shiyer
On 12/03/2012 06:14 PM, Michael S. Tsirkin wrote:
> On Tue, Nov 27, 2012 at 06:15:59PM +0800, Jason Wang wrote:
>> > - if (!try_fill_recv(&vi->rq, GFP_KERNEL))
>> > - schedule_delayed_work(&vi->rq.refill, 0);
>> > + for (i = 0; i < vi->max_queue_pairs; i++)
>> > + if (!try_fill_recv(&vi->rq[i], GFP_KERNEL))
>> > + schedule_delayed_work(&vi->rq[i].refill, 0);
>> >
>> > mutex_lock(&vi->config_lock);
>> > vi->config_enable = true;
>> > mutex_unlock(&vi->config_lock);
>> >
>> > + BUG_ON(virtnet_set_queues(vi));
>> > +
>> > return 0;
>> > }
>> > #endif
> Also crashing on device nack of command is also not nice.
> In this case it seems we can just switch to
> single-queue mode which should always be safe.
Not sure it's safe. It depends on the reason why this call fails. If we
left a state that the driver only use single queue but the device use
multi queues, we may still lost the network.
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 2/3] virtio_net: multiqueue support
2012-12-03 10:30 ` Jason Wang
@ 2012-12-03 11:00 ` Michael S. Tsirkin
2012-12-04 7:35 ` Michael S. Tsirkin
1 sibling, 0 replies; 26+ messages in thread
From: Michael S. Tsirkin @ 2012-12-03 11:00 UTC (permalink / raw)
To: Jason Wang
Cc: rusty, krkumar2, virtualization, netdev, linux-kernel, kvm,
bhutchings, jwhan, shiyer
On Mon, Dec 03, 2012 at 06:30:49PM +0800, Jason Wang wrote:
> On 12/03/2012 06:14 PM, Michael S. Tsirkin wrote:
> > On Tue, Nov 27, 2012 at 06:15:59PM +0800, Jason Wang wrote:
> >> > - if (!try_fill_recv(&vi->rq, GFP_KERNEL))
> >> > - schedule_delayed_work(&vi->rq.refill, 0);
> >> > + for (i = 0; i < vi->max_queue_pairs; i++)
> >> > + if (!try_fill_recv(&vi->rq[i], GFP_KERNEL))
> >> > + schedule_delayed_work(&vi->rq[i].refill, 0);
> >> >
> >> > mutex_lock(&vi->config_lock);
> >> > vi->config_enable = true;
> >> > mutex_unlock(&vi->config_lock);
> >> >
> >> > + BUG_ON(virtnet_set_queues(vi));
> >> > +
> >> > return 0;
> >> > }
> >> > #endif
> > Also crashing on device nack of command is also not nice.
> > In this case it seems we can just switch to
> > single-queue mode which should always be safe.
>
> Not sure it's safe. It depends on the reason why this call fails. If we
> left a state that the driver only use single queue but the device use
> multi queues, we may still lost the network.
Not the way driver is currently written - you'll happily
process incoming packets from all queues so no problem?
--
MST
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 2/3] virtio_net: multiqueue support
2012-12-03 10:01 ` Jason Wang
@ 2012-12-03 11:11 ` Michael S. Tsirkin
2012-12-04 9:24 ` Jason Wang
0 siblings, 1 reply; 26+ messages in thread
From: Michael S. Tsirkin @ 2012-12-03 11:11 UTC (permalink / raw)
To: Jason Wang
Cc: Rusty Russell, krkumar2, virtualization, netdev, linux-kernel,
kvm, bhutchings, jwhan, shiyer
On Mon, Dec 03, 2012 at 06:01:58PM +0800, Jason Wang wrote:
> On 12/03/2012 05:47 PM, Michael S. Tsirkin wrote:
> > On Mon, Dec 03, 2012 at 02:05:27PM +0800, Jason Wang wrote:
> >> On Monday, December 03, 2012 12:34:08 PM Rusty Russell wrote:
> >>> Jason Wang <jasowang@redhat.com> writes:
> >>>> +static const struct ethtool_ops virtnet_ethtool_ops;
> >>>> +
> >>>> +/*
> >>>> + * Converting between virtqueue no. and kernel tx/rx queue no.
> >>>> + * 0:rx0 1:tx0 2:cvq 3:rx1 4:tx1 ... 2N+1:rxN 2N+2:txN
> >>>> + */
> >>>> +static int vq2txq(struct virtqueue *vq)
> >>>> +{
> >>>> + int index = virtqueue_get_queue_index(vq);
> >>>> + return index == 1 ? 0 : (index - 2) / 2;
> >>>> +}
> >>>> +
> >>>> +static int txq2vq(int txq)
> >>>> +{
> >>>> + return txq ? 2 * txq + 2 : 1;
> >>>> +}
> >>>> +
> >>>> +static int vq2rxq(struct virtqueue *vq)
> >>>> +{
> >>>> + int index = virtqueue_get_queue_index(vq);
> >>>> + return index ? (index - 1) / 2 : 0;
> >>>> +}
> >>>> +
> >>>> +static int rxq2vq(int rxq)
> >>>> +{
> >>>> + return rxq ? 2 * rxq + 1 : 0;
> >>>> +}
> >>>> +
> >>> I thought MST changed the proposed spec to make the control queue always
> >>> the last one, so this logic becomes trivial.
> >> But it may break the support of legacy guest. If we boot a legacy single queue
> >> guest on a 2 queue virtio-net device. It may think vq 2 is cvq which is indeed
> >> rx1.
> > Legacy guyest support should be handled by host using feature
> > bits in the usual way: host should detect legacy guest
> > by checking the VIRTIO_NET_F_RFS feature.
> >
> > If VIRTIO_NET_F_RFS is acked, cvq is vq max_virtqueue_pairs * 2.
> > If it's not acked, cvq is vq 2.
> >
>
> We could, but we didn't gain much from this.
It just seems cleaner and easier to understand.
> Furthermore, we need also
> do the dynamic creation/destroying of virtqueues during feature
> negotiation which seems not supported in qemu now.
It's not *done* in qemu now, but it seems easy: just call
virtio_add_queue for vq2 and on from virtio_net_set_features.
As features can be modified multiple times, we
should add virtio_del_queue and call that beforehand
to get to the known state (two vqs).
--
MST
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 1/3] virtio-net: separate fields of sending/receiving queue from virtnet_info
[not found] ` <3524590.ZWGua7A8ne@jason-thinkpad-t430s>
@ 2012-12-03 11:18 ` Michael S. Tsirkin
2012-12-04 9:22 ` Jason Wang
2012-12-04 3:43 ` Rusty Russell
1 sibling, 1 reply; 26+ messages in thread
From: Michael S. Tsirkin @ 2012-12-03 11:18 UTC (permalink / raw)
To: Jason Wang
Cc: Rusty Russell, krkumar2, virtualization, netdev, linux-kernel,
kvm, bhutchings, jwhan, shiyer
On Mon, Dec 03, 2012 at 01:15:01PM +0800, Jason Wang wrote:
> > > +
>
> > > + /* Work struct for refilling if we run low on memory. */
>
> > > + struct delayed_work refill;
>
> >
>
> > I can't really see the justificaiton for a refill per queue. Just have
>
> > one work iterate all the queues if it happens, unless it happens often
>
> > (in which case, we need to look harder at this anyway).
>
>
>
> But during this kind of iteration, we may need enable/disable the napi
> regardless of whether the receive queue has lots to be refilled. This may add
> extra latency.
We are running from the timer, so latency is not a concern I think.
--
MST
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 3/3] virtio-net: change the number of queues through ethtool
2012-12-03 6:09 ` Jason Wang
@ 2012-12-03 11:25 ` Michael S. Tsirkin
2012-12-04 0:22 ` Ben Hutchings
0 siblings, 1 reply; 26+ messages in thread
From: Michael S. Tsirkin @ 2012-12-03 11:25 UTC (permalink / raw)
To: Jason Wang
Cc: rusty, krkumar2, virtualization, netdev, linux-kernel, kvm,
bhutchings, jwhan, shiyer
On Mon, Dec 03, 2012 at 02:09:28PM +0800, Jason Wang wrote:
> On Sunday, December 02, 2012 06:09:06 PM Michael S. Tsirkin wrote:
> > On Tue, Nov 27, 2012 at 06:16:00PM +0800, Jason Wang wrote:
> > > This patch implement the {set|get}_channels method of ethool to allow user
> > > to change the number of queues dymaically when the device is running.
> > > This would let the user to configure it on demand.
> > >
> > > Signed-off-by: Jason Wang <jasowang@redhat.com>
> > > ---
> > >
> > > drivers/net/virtio_net.c | 41 +++++++++++++++++++++++++++++++++++++++++
> > > 1 files changed, 41 insertions(+), 0 deletions(-)
> > >
> > > diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> > > index bcaa6e5..f08ec2a 100644
> > > --- a/drivers/net/virtio_net.c
> > > +++ b/drivers/net/virtio_net.c
> > > @@ -1578,10 +1578,51 @@ static struct virtio_driver virtio_net_driver = {
> > >
> > > #endif
> > > };
> > >
> > > +/* TODO: Eliminate OOO packets during switching */
> > > +static int virtnet_set_channels(struct net_device *dev,
> > > + struct ethtool_channels *channels)
> > > +{
> > > + struct virtnet_info *vi = netdev_priv(dev);
> > > + u16 queue_pairs = channels->combined_count;
by the way shouldn't this be combined_count / 2?
And below channels->max_combined = vi->max_queue_pairs * 2; ?
> > > +
> > > + /* We don't support separate rx/tx channels.
> > > + * We don't allow setting 'other' channels.
> > > + */
> > > + if (channels->rx_count || channels->tx_count || channels->other_count)
> > > + return -EINVAL;
> > > +
> > > + /* Only two modes were support currently */
> > > + if (queue_pairs != vi->max_queue_pairs && queue_pairs != 1)
> > > + return -EINVAL;
> > > +
> >
> > Why the limitation?
>
> Not sure the value bettwen 1 and max_queue_pairs is useful. But anyway, I can
> remove this limitation.
I guess a reasonable value would be number of active CPUs,
which can change with CPU hotplug.
> > Also how does userspace discover what the legal values are?
>
> Userspace only check whether the value is greater than max_queue_pairs.
Exactly. One can query max combined but the fact that configuring
any less is impossible might be surprising.
> >
> > > + vi->curr_queue_pairs = queue_pairs;
> > > + BUG_ON(virtnet_set_queues(vi));
> > > +
> > > + netif_set_real_num_tx_queues(dev, vi->curr_queue_pairs);
> > > + netif_set_real_num_rx_queues(dev, vi->curr_queue_pairs);
> > > +
> > > + return 0;
> > > +}
> > > +
> > > +static void virtnet_get_channels(struct net_device *dev,
> > > + struct ethtool_channels *channels)
> > > +{
> > > + struct virtnet_info *vi = netdev_priv(dev);
> > > +
> > > + channels->combined_count = vi->curr_queue_pairs;
> > > + channels->max_combined = vi->max_queue_pairs;
> > > + channels->max_other = 0;
> > > + channels->rx_count = 0;
> > > + channels->tx_count = 0;
> > > + channels->other_count = 0;
> > > +}
> > > +
> > >
> > > static const struct ethtool_ops virtnet_ethtool_ops = {
> > >
> > > .get_drvinfo = virtnet_get_drvinfo,
> > > .get_link = ethtool_op_get_link,
> > > .get_ringparam = virtnet_get_ringparam,
> > >
> > > + .set_channels = virtnet_set_channels,
> > > + .get_channels = virtnet_get_channels,
> > >
> > > };
> > >
> > > static int __init init(void)
> >
> > --
> > To unsubscribe from this list: send the line "unsubscribe kvm" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 3/3] virtio-net: change the number of queues through ethtool
2012-12-03 11:25 ` Michael S. Tsirkin
@ 2012-12-04 0:22 ` Ben Hutchings
0 siblings, 0 replies; 26+ messages in thread
From: Ben Hutchings @ 2012-12-04 0:22 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: Jason Wang, rusty, krkumar2, virtualization, netdev,
linux-kernel, kvm, jwhan, shiyer
On Mon, 2012-12-03 at 13:25 +0200, Michael S. Tsirkin wrote:
> On Mon, Dec 03, 2012 at 02:09:28PM +0800, Jason Wang wrote:
> > On Sunday, December 02, 2012 06:09:06 PM Michael S. Tsirkin wrote:
> > > On Tue, Nov 27, 2012 at 06:16:00PM +0800, Jason Wang wrote:
> > > > This patch implement the {set|get}_channels method of ethool to allow user
> > > > to change the number of queues dymaically when the device is running.
> > > > This would let the user to configure it on demand.
> > > >
> > > > Signed-off-by: Jason Wang <jasowang@redhat.com>
> > > > ---
> > > >
> > > > drivers/net/virtio_net.c | 41 +++++++++++++++++++++++++++++++++++++++++
> > > > 1 files changed, 41 insertions(+), 0 deletions(-)
> > > >
> > > > diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> > > > index bcaa6e5..f08ec2a 100644
> > > > --- a/drivers/net/virtio_net.c
> > > > +++ b/drivers/net/virtio_net.c
> > > > @@ -1578,10 +1578,51 @@ static struct virtio_driver virtio_net_driver = {
> > > >
> > > > #endif
> > > > };
> > > >
> > > > +/* TODO: Eliminate OOO packets during switching */
> > > > +static int virtnet_set_channels(struct net_device *dev,
> > > > + struct ethtool_channels *channels)
> > > > +{
> > > > + struct virtnet_info *vi = netdev_priv(dev);
> > > > + u16 queue_pairs = channels->combined_count;
>
> by the way shouldn't this be combined_count / 2?
>
> And below channels->max_combined = vi->max_queue_pairs * 2; ?
[...]
In this ethtool API, 'channel' means an IRQ and set of queues that
trigger it. So each ethtool-channel will correspond to one queue-pair
and not one virtio channel.
Ben.
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 1/3] virtio-net: separate fields of sending/receiving queue from virtnet_info
[not found] ` <3524590.ZWGua7A8ne@jason-thinkpad-t430s>
2012-12-03 11:18 ` Michael S. Tsirkin
@ 2012-12-04 3:43 ` Rusty Russell
2012-12-04 9:23 ` Jason Wang
1 sibling, 1 reply; 26+ messages in thread
From: Rusty Russell @ 2012-12-04 3:43 UTC (permalink / raw)
To: Jason Wang
Cc: mst, krkumar2, virtualization, netdev, linux-kernel, kvm,
bhutchings, jwhan, shiyer
Jason Wang <jasowang@redhat.com> writes:
> On Monday, December 03, 2012 12:25:42 PM Rusty Russell wrote:
>> > +
>> > + /* Work struct for refilling if we run low on memory. */
>> > + struct delayed_work refill;
>>
>> I can't really see the justificaiton for a refill per queue. Just have
>> one work iterate all the queues if it happens, unless it happens often
>> (in which case, we need to look harder at this anyway).
>
> But during this kind of iteration, we may need enable/disable the napi
> regardless of whether the receive queue has lots to be refilled. This may add
> extra latency.
Sure, but does it actually happen? We only use the work when we run out
of memory. If this happens in normal behaviour we need to change
something else...
Thanks,
Rusty.
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 2/3] virtio_net: multiqueue support
2012-12-03 10:30 ` Jason Wang
2012-12-03 11:00 ` Michael S. Tsirkin
@ 2012-12-04 7:35 ` Michael S. Tsirkin
2012-12-04 9:27 ` Jason Wang
1 sibling, 1 reply; 26+ messages in thread
From: Michael S. Tsirkin @ 2012-12-04 7:35 UTC (permalink / raw)
To: Jason Wang
Cc: rusty, krkumar2, virtualization, netdev, linux-kernel, kvm,
bhutchings, jwhan, shiyer
On Mon, Dec 03, 2012 at 06:30:49PM +0800, Jason Wang wrote:
> On 12/03/2012 06:14 PM, Michael S. Tsirkin wrote:
> > On Tue, Nov 27, 2012 at 06:15:59PM +0800, Jason Wang wrote:
> >> > - if (!try_fill_recv(&vi->rq, GFP_KERNEL))
> >> > - schedule_delayed_work(&vi->rq.refill, 0);
> >> > + for (i = 0; i < vi->max_queue_pairs; i++)
> >> > + if (!try_fill_recv(&vi->rq[i], GFP_KERNEL))
> >> > + schedule_delayed_work(&vi->rq[i].refill, 0);
> >> >
> >> > mutex_lock(&vi->config_lock);
> >> > vi->config_enable = true;
> >> > mutex_unlock(&vi->config_lock);
> >> >
> >> > + BUG_ON(virtnet_set_queues(vi));
> >> > +
> >> > return 0;
> >> > }
> >> > #endif
> > Also crashing on device nack of command is also not nice.
> > In this case it seems we can just switch to
> > single-queue mode which should always be safe.
>
> Not sure it's safe. It depends on the reason why this call fails. If we
> left a state that the driver only use single queue but the device use
> multi queues, we may still lost the network.
Looks like we won't: napi will stay enabled on all queues
so we will process incoming packets.
--
MST
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 1/3] virtio-net: separate fields of sending/receiving queue from virtnet_info
2012-12-03 11:18 ` Michael S. Tsirkin
@ 2012-12-04 9:22 ` Jason Wang
0 siblings, 0 replies; 26+ messages in thread
From: Jason Wang @ 2012-12-04 9:22 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: Rusty Russell, krkumar2, virtualization, netdev, linux-kernel,
kvm, bhutchings, jwhan, shiyer
On Monday, December 03, 2012 01:18:48 PM Michael S. Tsirkin wrote:
> On Mon, Dec 03, 2012 at 01:15:01PM +0800, Jason Wang wrote:
> > > > +
> > > >
> > > > + /* Work struct for refilling if we run low on memory. */
> > > >
> > > > + struct delayed_work refill;
> > >
> > > I can't really see the justificaiton for a refill per queue. Just have
> > >
> > > one work iterate all the queues if it happens, unless it happens often
> > >
> > > (in which case, we need to look harder at this anyway).
> >
> > But during this kind of iteration, we may need enable/disable the napi
> > regardless of whether the receive queue has lots to be refilled. This may
> > add extra latency.
>
> We are running from the timer, so latency is not a concern I think.
Maybe, anyway it's only called when avaiable memory is low, so it should not
be an issue.
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 1/3] virtio-net: separate fields of sending/receiving queue from virtnet_info
2012-12-04 3:43 ` Rusty Russell
@ 2012-12-04 9:23 ` Jason Wang
0 siblings, 0 replies; 26+ messages in thread
From: Jason Wang @ 2012-12-04 9:23 UTC (permalink / raw)
To: Rusty Russell
Cc: mst, krkumar2, virtualization, netdev, linux-kernel, kvm,
bhutchings, jwhan, shiyer
On Tuesday, December 04, 2012 02:13:11 PM Rusty Russell wrote:
> Jason Wang <jasowang@redhat.com> writes:
> > On Monday, December 03, 2012 12:25:42 PM Rusty Russell wrote:
> >> > +
> >> > + /* Work struct for refilling if we run low on memory. */
> >> > + struct delayed_work refill;
> >>
> >> I can't really see the justificaiton for a refill per queue. Just have
> >> one work iterate all the queues if it happens, unless it happens often
> >> (in which case, we need to look harder at this anyway).
> >
> > But during this kind of iteration, we may need enable/disable the napi
> > regardless of whether the receive queue has lots to be refilled. This may
> > add extra latency.
>
> Sure, but does it actually happen? We only use the work when we run out
> of memory. If this happens in normal behaviour we need to change
> something else...
True, I will change to use a global one.
Thanks
>
> Thanks,
> Rusty.
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 2/3] virtio_net: multiqueue support
2012-12-03 11:11 ` Michael S. Tsirkin
@ 2012-12-04 9:24 ` Jason Wang
0 siblings, 0 replies; 26+ messages in thread
From: Jason Wang @ 2012-12-04 9:24 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: Rusty Russell, krkumar2, virtualization, netdev, linux-kernel,
kvm, bhutchings, jwhan, shiyer
On Monday, December 03, 2012 01:11:18 PM Michael S. Tsirkin wrote:
> On Mon, Dec 03, 2012 at 06:01:58PM +0800, Jason Wang wrote:
> > On 12/03/2012 05:47 PM, Michael S. Tsirkin wrote:
> > > On Mon, Dec 03, 2012 at 02:05:27PM +0800, Jason Wang wrote:
> > >> On Monday, December 03, 2012 12:34:08 PM Rusty Russell wrote:
> > >>> Jason Wang <jasowang@redhat.com> writes:
> > >>>> +static const struct ethtool_ops virtnet_ethtool_ops;
> > >>>> +
> > >>>> +/*
> > >>>> + * Converting between virtqueue no. and kernel tx/rx queue no.
> > >>>> + * 0:rx0 1:tx0 2:cvq 3:rx1 4:tx1 ... 2N+1:rxN 2N+2:txN
> > >>>> + */
> > >>>> +static int vq2txq(struct virtqueue *vq)
> > >>>> +{
> > >>>> + int index = virtqueue_get_queue_index(vq);
> > >>>> + return index == 1 ? 0 : (index - 2) / 2;
> > >>>> +}
> > >>>> +
> > >>>> +static int txq2vq(int txq)
> > >>>> +{
> > >>>> + return txq ? 2 * txq + 2 : 1;
> > >>>> +}
> > >>>> +
> > >>>> +static int vq2rxq(struct virtqueue *vq)
> > >>>> +{
> > >>>> + int index = virtqueue_get_queue_index(vq);
> > >>>> + return index ? (index - 1) / 2 : 0;
> > >>>> +}
> > >>>> +
> > >>>> +static int rxq2vq(int rxq)
> > >>>> +{
> > >>>> + return rxq ? 2 * rxq + 1 : 0;
> > >>>> +}
> > >>>> +
> > >>>
> > >>> I thought MST changed the proposed spec to make the control queue
> > >>> always
> > >>> the last one, so this logic becomes trivial.
> > >>
> > >> But it may break the support of legacy guest. If we boot a legacy
> > >> single queue guest on a 2 queue virtio-net device. It may think vq 2
> > >> is cvq which is indeed rx1.
> > >
> > > Legacy guyest support should be handled by host using feature
> > > bits in the usual way: host should detect legacy guest
> > > by checking the VIRTIO_NET_F_RFS feature.
> > >
> > > If VIRTIO_NET_F_RFS is acked, cvq is vq max_virtqueue_pairs * 2.
> > > If it's not acked, cvq is vq 2.
> >
> > We could, but we didn't gain much from this.
>
> It just seems cleaner and easier to understand.
>
> > Furthermore, we need also
> > do the dynamic creation/destroying of virtqueues during feature
> > negotiation which seems not supported in qemu now.
>
> It's not *done* in qemu now, but it seems easy: just call
> virtio_add_queue for vq2 and on from virtio_net_set_features.
> As features can be modified multiple times, we
> should add virtio_del_queue and call that beforehand
> to get to the known state (two vqs).
And also need some work after migration like what we need in setting features.
I'm ok with this method, will change to follow spec v5.
Thanks
^ permalink raw reply [flat|nested] 26+ messages in thread
* Re: [net-next rfc v7 2/3] virtio_net: multiqueue support
2012-12-04 7:35 ` Michael S. Tsirkin
@ 2012-12-04 9:27 ` Jason Wang
0 siblings, 0 replies; 26+ messages in thread
From: Jason Wang @ 2012-12-04 9:27 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: rusty, krkumar2, virtualization, netdev, linux-kernel, kvm,
bhutchings, jwhan, shiyer
On Tuesday, December 04, 2012 09:35:03 AM Michael S. Tsirkin wrote:
> On Mon, Dec 03, 2012 at 06:30:49PM +0800, Jason Wang wrote:
> > On 12/03/2012 06:14 PM, Michael S. Tsirkin wrote:
> > > On Tue, Nov 27, 2012 at 06:15:59PM +0800, Jason Wang wrote:
> > >> > - if (!try_fill_recv(&vi->rq, GFP_KERNEL))
> > >> > - schedule_delayed_work(&vi->rq.refill, 0);
> > >> > + for (i = 0; i < vi->max_queue_pairs; i++)
> > >> > + if (!try_fill_recv(&vi->rq[i], GFP_KERNEL))
> > >> > + schedule_delayed_work(&vi->rq[i].refill, 0);
> > >> >
> > >> > mutex_lock(&vi->config_lock);
> > >> > vi->config_enable = true;
> > >> > mutex_unlock(&vi->config_lock);
> > >> >
> > >> > + BUG_ON(virtnet_set_queues(vi));
> > >> > +
> > >> >
> > >> > return 0;
> > >> >
> > >> > }
> > >> > #endif
> > >
> > > Also crashing on device nack of command is also not nice.
> > > In this case it seems we can just switch to
> > > single-queue mode which should always be safe.
> >
> > Not sure it's safe. It depends on the reason why this call fails. If we
> > left a state that the driver only use single queue but the device use
> > multi queues, we may still lost the network.
>
> Looks like we won't: napi will stay enabled on all queues
> so we will process incoming packets.
True, consider there's no bug in qemu. Will leave a just leave a warning in
next version.
Thanks
^ permalink raw reply [flat|nested] 26+ messages in thread
end of thread, other threads:[~2012-12-04 9:27 UTC | newest]
Thread overview: 26+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2012-11-27 10:15 [net-next rfc v7 0/3] Multiqueue virtio-net Jason Wang
2012-11-27 10:15 ` [net-next rfc v7 1/3] virtio-net: separate fields of sending/receiving queue from virtnet_info Jason Wang
2012-12-03 1:55 ` Rusty Russell
[not found] ` <3524590.ZWGua7A8ne@jason-thinkpad-t430s>
2012-12-03 11:18 ` Michael S. Tsirkin
2012-12-04 9:22 ` Jason Wang
2012-12-04 3:43 ` Rusty Russell
2012-12-04 9:23 ` Jason Wang
2012-11-27 10:15 ` [net-next rfc v7 2/3] virtio_net: multiqueue support Jason Wang
2012-12-02 16:06 ` Michael S. Tsirkin
2012-12-03 5:47 ` Jason Wang
2012-12-03 2:04 ` Rusty Russell
2012-12-03 6:05 ` Jason Wang
2012-12-03 9:47 ` Michael S. Tsirkin
2012-12-03 10:01 ` Jason Wang
2012-12-03 11:11 ` Michael S. Tsirkin
2012-12-04 9:24 ` Jason Wang
2012-12-03 10:14 ` Michael S. Tsirkin
2012-12-03 10:30 ` Jason Wang
2012-12-03 11:00 ` Michael S. Tsirkin
2012-12-04 7:35 ` Michael S. Tsirkin
2012-12-04 9:27 ` Jason Wang
2012-11-27 10:16 ` [net-next rfc v7 3/3] virtio-net: change the number of queues through ethtool Jason Wang
2012-12-02 16:09 ` Michael S. Tsirkin
2012-12-03 6:09 ` Jason Wang
2012-12-03 11:25 ` Michael S. Tsirkin
2012-12-04 0:22 ` Ben Hutchings
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).