DPDK patches and discussions
 help / color / mirror / Atom feed
From: Matthias Gatto <matthias.gatto@outscale.com>
To: dev@dpdk.org
Cc: Matthias Gatto <matthias.gatto@outscale.com>,
	maxime.coquelin@redhat.com, tiwei.bie@intel.com,
	zhihong.wang@intel.com
Subject: [dpdk-dev] [PATCH] vhost: fix race condition in fdset_add
Date: Thu,  6 Dec 2018 16:00:07 +0000	[thread overview]
Message-ID: <1544112007-23177-1-git-send-email-matthias.gatto@outscale.com> (raw)

fdset_add can call fdset_shrink_nolock which call fdset_move
concurrently to poll that is call in fdset_event_dispatch.

This patch add a mutex to protect poll from been call at the same time
fdset_add call fdset_shrink_nolock.

Signed-off-by: Matthias Gatto <matthias.gatto@outscale.com>
---
 lib/librte_vhost/fd_man.c | 4 ++++
 lib/librte_vhost/fd_man.h | 1 +
 lib/librte_vhost/socket.c | 1 +
 3 files changed, 6 insertions(+)

diff --git a/lib/librte_vhost/fd_man.c b/lib/librte_vhost/fd_man.c
index 38347ab..55d4856 100644
--- a/lib/librte_vhost/fd_man.c
+++ b/lib/librte_vhost/fd_man.c
@@ -129,7 +129,9 @@
 	pthread_mutex_lock(&pfdset->fd_mutex);
 	i = pfdset->num < MAX_FDS ? pfdset->num++ : -1;
 	if (i == -1) {
+		pthread_mutex_lock(&pfdset->fd_pooling_mutex);
 		fdset_shrink_nolock(pfdset);
+		pthread_mutex_unlock(&pfdset->fd_pooling_mutex);
 		i = pfdset->num < MAX_FDS ? pfdset->num++ : -1;
 		if (i == -1) {
 			pthread_mutex_unlock(&pfdset->fd_mutex);
@@ -246,7 +248,9 @@
 		numfds = pfdset->num;
 		pthread_mutex_unlock(&pfdset->fd_mutex);
 
+		pthread_mutex_lock(&pfdset->fd_pooling_mutex);
 		val = poll(pfdset->rwfds, numfds, 1000 /* millisecs */);
+		pthread_mutex_unlock(&pfdset->fd_pooling_mutex);
 		if (val < 0)
 			continue;
 
diff --git a/lib/librte_vhost/fd_man.h b/lib/librte_vhost/fd_man.h
index 3331bcd..3ab5cfd 100644
--- a/lib/librte_vhost/fd_man.h
+++ b/lib/librte_vhost/fd_man.h
@@ -24,6 +24,7 @@ struct fdset {
 	struct pollfd rwfds[MAX_FDS];
 	struct fdentry fd[MAX_FDS];
 	pthread_mutex_t fd_mutex;
+	pthread_mutex_t fd_pooling_mutex;
 	int num;	/* current fd number of this fdset */
 
 	union pipefds {
diff --git a/lib/librte_vhost/socket.c b/lib/librte_vhost/socket.c
index d630317..cc4e748 100644
--- a/lib/librte_vhost/socket.c
+++ b/lib/librte_vhost/socket.c
@@ -88,6 +88,7 @@ struct vhost_user {
 	.fdset = {
 		.fd = { [0 ... MAX_FDS - 1] = {-1, NULL, NULL, NULL, 0} },
 		.fd_mutex = PTHREAD_MUTEX_INITIALIZER,
+		.fd_pooling_mutex = PTHREAD_MUTEX_INITIALIZER,
 		.num = 0
 	},
 	.vsocket_cnt = 0,
-- 
1.8.3.1

             reply	other threads:[~2018-12-06 16:00 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-12-06 16:00 Matthias Gatto [this message]
2018-12-11 18:11 ` Maxime Coquelin
2018-12-14  9:32   ` Matthias Gatto
2018-12-14  9:51     ` Maxime Coquelin
2018-12-14  9:53       ` Maxime Coquelin
2018-12-14 10:07         ` Matthias Gatto
2018-12-14 10:08           ` Maxime Coquelin
2018-12-18 14:01 ` 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=1544112007-23177-1-git-send-email-matthias.gatto@outscale.com \
    --to=matthias.gatto@outscale.com \
    --cc=dev@dpdk.org \
    --cc=maxime.coquelin@redhat.com \
    --cc=tiwei.bie@intel.com \
    --cc=zhihong.wang@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).