From: Heng Qi <hengqi@linux.alibaba.com>
To: netdev@vger.kernel.org, virtualization@lists.linux.dev
Cc: "Michael S . Tsirkin" <mst@redhat.com>,
Jason Wang <jasowang@redhat.com>,
Xuan Zhuo <xuanzhuo@linux.alibaba.com>,
Eric Dumazet <edumazet@google.com>,
"David S . Miller" <davem@davemloft.net>,
Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>
Subject: [PATCH net-next 1/3] virtio_net: enable irq for the control vq
Date: Thu, 25 Apr 2024 20:58:53 +0800 [thread overview]
Message-ID: <20240425125855.87025-2-hengqi@linux.alibaba.com> (raw)
In-Reply-To: <20240425125855.87025-1-hengqi@linux.alibaba.com>
Control vq polling request results consume more CPU.
Especially when dim issues more control requests to the device,
it's beneficial to the guest to enable control vq's irq.
Suggested-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Heng Qi <hengqi@linux.alibaba.com>
---
drivers/net/virtio_net.c | 45 ++++++++++++++++++++++++++++++----------
1 file changed, 34 insertions(+), 11 deletions(-)
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index a4d3c76654a4..79a1b30c173c 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -287,6 +287,12 @@ struct virtnet_info {
bool has_cvq;
struct mutex cvq_lock;
+ /* Wait for the device to complete the request */
+ struct completion completion;
+
+ /* Work struct for acquisition of cvq processing results. */
+ struct work_struct get_cvq;
+
/* Host can handle any s/g split between our header and packet data */
bool any_header_sg;
@@ -520,6 +526,13 @@ static bool virtqueue_napi_complete(struct napi_struct *napi,
return false;
}
+static void virtnet_cvq_done(struct virtqueue *cvq)
+{
+ struct virtnet_info *vi = cvq->vdev->priv;
+
+ schedule_work(&vi->get_cvq);
+}
+
static void skb_xmit_done(struct virtqueue *vq)
{
struct virtnet_info *vi = vq->vdev->priv;
@@ -2036,6 +2049,20 @@ static bool try_fill_recv(struct virtnet_info *vi, struct receive_queue *rq,
return !oom;
}
+static void virtnet_get_cvq_work(struct work_struct *work)
+{
+ struct virtnet_info *vi =
+ container_of(work, struct virtnet_info, get_cvq);
+ unsigned int tmp;
+ void *res;
+
+ mutex_lock(&vi->cvq_lock);
+ res = virtqueue_get_buf(vi->cvq, &tmp);
+ if (res)
+ complete(&vi->completion);
+ mutex_unlock(&vi->cvq_lock);
+}
+
static void skb_recv_done(struct virtqueue *rvq)
{
struct virtnet_info *vi = rvq->vdev->priv;
@@ -2531,7 +2558,7 @@ static bool virtnet_send_command(struct virtnet_info *vi, u8 class, u8 cmd,
struct scatterlist *out)
{
struct scatterlist *sgs[4], hdr, stat;
- unsigned out_num = 0, tmp;
+ unsigned out_num = 0;
int ret;
/* Caller should know better */
@@ -2566,16 +2593,10 @@ static bool virtnet_send_command(struct virtnet_info *vi, u8 class, u8 cmd,
return vi->ctrl->status == VIRTIO_NET_OK;
}
- /* Spin for a response, the kick causes an ioport write, trapping
- * into the hypervisor, so the request should be handled immediately.
- */
- while (!virtqueue_get_buf(vi->cvq, &tmp) &&
- !virtqueue_is_broken(vi->cvq)) {
- cond_resched();
- cpu_relax();
- }
-
mutex_unlock(&vi->cvq_lock);
+
+ wait_for_completion(&vi->completion);
+
return vi->ctrl->status == VIRTIO_NET_OK;
}
@@ -4433,7 +4454,7 @@ static int virtnet_find_vqs(struct virtnet_info *vi)
/* Parameters for control virtqueue, if any */
if (vi->has_cvq) {
- callbacks[total_vqs - 1] = NULL;
+ callbacks[total_vqs - 1] = virtnet_cvq_done;
names[total_vqs - 1] = "control";
}
@@ -4952,6 +4973,8 @@ static int virtnet_probe(struct virtio_device *vdev)
if (vi->has_rss || vi->has_rss_hash_report)
virtnet_init_default_rss(vi);
+ INIT_WORK(&vi->get_cvq, virtnet_get_cvq_work);
+ init_completion(&vi->completion);
enable_rx_mode_work(vi);
/* serialize netdev register + virtio_device_ready() with ndo_open() */
--
2.32.0.3.g01195cf9f
next prev parent reply other threads:[~2024-04-25 12:59 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-25 12:58 [PATCH net-next 0/3] virtio_net: enable the irq for ctrlq Heng Qi
2024-04-25 12:58 ` Heng Qi [this message]
2024-05-07 3:15 ` [PATCH net-next 1/3] virtio_net: enable irq for the control vq Jason Wang
2024-05-07 3:55 ` Heng Qi
2024-05-07 6:24 ` Jason Wang
2024-05-07 6:27 ` Heng Qi
2024-05-08 2:19 ` Jason Wang
2024-05-08 3:43 ` Heng Qi
2024-04-25 12:58 ` [PATCH net-next 2/3] virtio_net: fix possible dim status unrecoverable Heng Qi
2024-04-26 2:21 ` Jakub Kicinski
2024-04-26 2:55 ` Heng Qi
2024-04-25 12:58 ` [PATCH net-next 3/3] virtio_net: improve dim command request efficiency Heng Qi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240425125855.87025-2-hengqi@linux.alibaba.com \
--to=hengqi@linux.alibaba.com \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=jasowang@redhat.com \
--cc=kuba@kernel.org \
--cc=mst@redhat.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=virtualization@lists.linux.dev \
--cc=xuanzhuo@linux.alibaba.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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 read-only IMAP folder(s) and NNTP newsgroup(s).