From: xiangxia.m.yue@gmail.com
To: jianfeng.tan@intel.com
Cc: dev@dpdk.org, Tonghao Zhang <xiangxia.m.yue@gmail.com>
Subject: [dpdk-dev] [PATCH 2/2] vhost: add pipe event for optimizing negotiating
Date: Tue, 27 Mar 2018 22:49:25 -0700 [thread overview]
Message-ID: <1522216165-19666-3-git-send-email-xiangxia.m.yue@gmail.com> (raw)
In-Reply-To: <1522216165-19666-1-git-send-email-xiangxia.m.yue@gmail.com>
From: Tonghao Zhang <xiangxia.m.yue@gmail.com>
When vhost—user connects qemu successfully, dpdk will call
the vhost_user_add_connection to add unix socket fd to poll.
And fdset_add only set the socket fd to a fdentry while poll
may sleep now. In a general case, this is no problem. But if
we use hot update for vhost-user, most downtime of VMs network
is 750+ms. This patch adds pipe event, so after connections are
ok, dpdk rebuild the poll immediately. With this patch, the
most downtime is 20~30ms.
Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com>
---
lib/librte_vhost/fd_man.c | 49 +++++++++++++++++++++++++++++++++++++++++++++++
lib/librte_vhost/fd_man.h | 16 ++++++++++++++++
lib/librte_vhost/socket.c | 14 ++++++++++++++
3 files changed, 79 insertions(+)
diff --git a/lib/librte_vhost/fd_man.c b/lib/librte_vhost/fd_man.c
index 181711c..7716757 100644
--- a/lib/librte_vhost/fd_man.c
+++ b/lib/librte_vhost/fd_man.c
@@ -15,6 +15,7 @@
#include <rte_log.h>
#include "fd_man.h"
+#include "vhost.h"
#define FDPOLLERR (POLLERR | POLLHUP | POLLNVAL)
@@ -272,3 +273,51 @@
return NULL;
}
+
+static void
+fdset_pipe_read_cb(int readfd, void *dat __rte_unused,
+ int *remove __rte_unused)
+{
+ char charbuf[16];
+ read(readfd, charbuf, sizeof(charbuf));
+}
+
+void
+fdset_pipe_uninit(struct fdset *fdset)
+{
+ fdset_del(fdset, fdset->u.readfd);
+ close(fdset->u.readfd);
+ close(fdset->u.writefd);
+}
+
+int
+fdset_pipe_init(struct fdset *fdset)
+{
+ int ret;
+
+ if (pipe(fdset->u.pipefd) < 0) {
+ RTE_LOG(ERR, VHOST_CONFIG,
+ "failed to create pipe for vhost fdset\n");
+ return -1;
+ }
+
+ ret = fdset_add(fdset, fdset->u.readfd,
+ fdset_pipe_read_cb, NULL, NULL);
+
+ if (ret < 0) {
+ RTE_LOG(ERR, VHOST_CONFIG,
+ "failed to add pipe readfd %d into vhost server fdset\n",
+ fdset->u.readfd);
+
+ fdset_pipe_uninit(fdset);
+ return -1;
+ }
+
+ return 0;
+}
+
+void
+fdset_pipe_notify(struct fdset *fdset)
+{
+ write(fdset->u.writefd, "1", 1);
+}
diff --git a/lib/librte_vhost/fd_man.h b/lib/librte_vhost/fd_man.h
index 3a9276c..76a42fb 100644
--- a/lib/librte_vhost/fd_man.h
+++ b/lib/librte_vhost/fd_man.h
@@ -25,6 +25,16 @@ struct fdset {
struct fdentry fd[MAX_FDS];
pthread_mutex_t fd_mutex;
int num; /* current fd number of this fdset */
+
+ union pipefds {
+ struct {
+ int pipefd[2];
+ };
+ struct {
+ int readfd;
+ int writefd;
+ };
+ } u;
};
@@ -37,4 +47,10 @@ int fdset_add(struct fdset *pfdset, int fd,
void *fdset_event_dispatch(void *arg);
+int fdset_pipe_init(struct fdset *fdset);
+
+void fdset_pipe_uninit(struct fdset *fdset);
+
+void fdset_pipe_notify(struct fdset *fdset);
+
#endif
diff --git a/lib/librte_vhost/socket.c b/lib/librte_vhost/socket.c
index 95bed78..795239c 100644
--- a/lib/librte_vhost/socket.c
+++ b/lib/librte_vhost/socket.c
@@ -231,6 +231,8 @@ struct vhost_user {
pthread_mutex_lock(&vsocket->conn_mutex);
TAILQ_INSERT_TAIL(&vsocket->conn_list, conn, next);
pthread_mutex_unlock(&vsocket->conn_mutex);
+
+ fdset_pipe_notify(&vhost_user.fdset);
return;
err:
@@ -829,11 +831,23 @@ struct vhost_device_ops const *
return -1;
if (fdset_tid == 0) {
+ /**
+ * create a pipe which will be waited by poll and notified to
+ * rebuild the wait list of poll.
+ */
+ if (fdset_pipe_init(&vhost_user.fdset) < 0) {
+ RTE_LOG(ERR, VHOST_CONFIG,
+ "failed to create pipe for vhost fdset\n");
+ return -1;
+ }
+
int ret = pthread_create(&fdset_tid, NULL, fdset_event_dispatch,
&vhost_user.fdset);
if (ret != 0) {
RTE_LOG(ERR, VHOST_CONFIG,
"failed to create fdset handling thread");
+
+ fdset_pipe_uninit(&vhost_user.fdset);
return -1;
} else {
snprintf(thread_name, RTE_MAX_THREAD_NAME_LEN,
--
1.8.3.1
next prev parent reply other threads:[~2018-03-28 5:49 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-03-28 5:49 [dpdk-dev] [PATCH 0/2] " xiangxia.m.yue
2018-03-28 5:49 ` [dpdk-dev] [PATCH 1/2] vhost: move stdbool header file to vhost.h xiangxia.m.yue
2018-03-30 8:09 ` Maxime Coquelin
2018-03-28 5:49 ` xiangxia.m.yue [this message]
2018-03-29 7:32 ` [dpdk-dev] [PATCH 2/2] vhost: add pipe event for optimizing negotiating Tan, Jianfeng
2018-03-29 11:24 ` Tonghao Zhang
2018-03-30 7:57 ` Maxime Coquelin
2018-03-30 8:07 ` Tonghao Zhang
2018-03-30 8:09 ` Maxime Coquelin
2018-03-30 8:17 ` [dpdk-dev] [PATCH 0/2] " Maxime Coquelin
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=1522216165-19666-3-git-send-email-xiangxia.m.yue@gmail.com \
--to=xiangxia.m.yue@gmail.com \
--cc=dev@dpdk.org \
--cc=jianfeng.tan@intel.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 NNTP newsgroup(s).