VPP-574: fix VPP hang during security group configuration on a suspended VM
The unix connect() in vhost-user driver in VPP is blocking, and
a non-expedient accept() on the other side causes the entire VPP to hang.
Solution: set the nonblocking flag for the socket fd before calling
connect(), and set the socket back to blocking after the accept() succeeds.
Change-Id: Ia5ee782037eeffabdad71db8241f476a048a4f6f
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
diff --git a/src/vnet/devices/virtio/vhost-user.c b/src/vnet/devices/virtio/vhost-user.c
index f9bbae4..3fbcee9 100644
--- a/src/vnet/devices/virtio/vhost-user.c
+++ b/src/vnet/devices/virtio/vhost-user.c
@@ -2325,9 +2325,13 @@
strncpy (sun.sun_path, (char *) vui->sock_filename,
sizeof (sun.sun_path) - 1);
+ /* Avoid hanging VPP if the other end does not accept */
+ fcntl(sockfd, F_SETFL, O_NONBLOCK);
if (connect (sockfd, (struct sockaddr *) &sun,
sizeof (struct sockaddr_un)) == 0)
{
+ /* Set the socket to blocking as it was before */
+ fcntl(sockfd, F_SETFL, 0);
vui->sock_errno = 0;
template.file_descriptor = sockfd;
template.private_data =