[20.02,v2] net/virtio-user: do not close tap when disabling queue pairs
Checks
Commit Message
Do not close the tap fds when disabling queue pairs, instead,
we just need to unbind the backend. Otherwise, tap port can be
destroyed unexpectedly.
Fixes: e3b434818bbb ("net/virtio-user: support kernel vhost")
Cc: stable@dpdk.org
Reported-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: Tiwei Bie <tiwei.bie@intel.com>
---
v2:
- Squash attach_queue/detach_queue into one routine (Stephen);
- Use initializer instead of memset (Stephen);
- Drop the cast to void * (Stephen);
drivers/net/virtio/virtio_user/vhost_kernel.c | 34 +++++++++++++++----
.../net/virtio/virtio_user/vhost_kernel_tap.c | 34 ++++++++++++-------
.../net/virtio/virtio_user/vhost_kernel_tap.h | 8 +++++
drivers/net/virtio/virtio_user/vhost_user.c | 4 +++
.../net/virtio/virtio_user/virtio_user_dev.c | 5 ++-
.../net/virtio/virtio_user/virtio_user_dev.h | 1 +
6 files changed, 65 insertions(+), 21 deletions(-)
Comments
On 11/25/19 9:14 AM, Tiwei Bie wrote:
> Do not close the tap fds when disabling queue pairs, instead,
> we just need to unbind the backend. Otherwise, tap port can be
> destroyed unexpectedly.
>
> Fixes: e3b434818bbb ("net/virtio-user: support kernel vhost")
> Cc: stable@dpdk.org
>
> Reported-by: Stephen Hemminger <stephen@networkplumber.org>
> Signed-off-by: Tiwei Bie <tiwei.bie@intel.com>
> ---
> v2:
> - Squash attach_queue/detach_queue into one routine (Stephen);
> - Use initializer instead of memset (Stephen);
> - Drop the cast to void * (Stephen);
>
> drivers/net/virtio/virtio_user/vhost_kernel.c | 34 +++++++++++++++----
> .../net/virtio/virtio_user/vhost_kernel_tap.c | 34 ++++++++++++-------
> .../net/virtio/virtio_user/vhost_kernel_tap.h | 8 +++++
> drivers/net/virtio/virtio_user/vhost_user.c | 4 +++
> .../net/virtio/virtio_user/virtio_user_dev.c | 5 ++-
> .../net/virtio/virtio_user/virtio_user_dev.h | 1 +
> 6 files changed, 65 insertions(+), 21 deletions(-)
>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Thanks,
Maxime
On 11/25/19 9:14 AM, Tiwei Bie wrote:
> Do not close the tap fds when disabling queue pairs, instead,
> we just need to unbind the backend. Otherwise, tap port can be
> destroyed unexpectedly.
>
> Fixes: e3b434818bbb ("net/virtio-user: support kernel vhost")
> Cc: stable@dpdk.org
>
> Reported-by: Stephen Hemminger <stephen@networkplumber.org>
> Signed-off-by: Tiwei Bie <tiwei.bie@intel.com>
> ---
> v2:
> - Squash attach_queue/detach_queue into one routine (Stephen);
> - Use initializer instead of memset (Stephen);
> - Drop the cast to void * (Stephen);
>
> drivers/net/virtio/virtio_user/vhost_kernel.c | 34 +++++++++++++++----
> .../net/virtio/virtio_user/vhost_kernel_tap.c | 34 ++++++++++++-------
> .../net/virtio/virtio_user/vhost_kernel_tap.h | 8 +++++
> drivers/net/virtio/virtio_user/vhost_user.c | 4 +++
> .../net/virtio/virtio_user/virtio_user_dev.c | 5 ++-
> .../net/virtio/virtio_user/virtio_user_dev.h | 1 +
> 6 files changed, 65 insertions(+), 21 deletions(-)
Applied to dpdk-next-virtio/master
Thanks,
Maxime
@@ -330,16 +330,34 @@ vhost_kernel_enable_queue_pair(struct virtio_user_dev *dev,
vhostfd = dev->vhostfds[pair_idx];
+ if (dev->qp_enabled[pair_idx] == enable)
+ return 0;
+
if (!enable) {
- if (dev->tapfds[pair_idx] >= 0) {
- close(dev->tapfds[pair_idx]);
- dev->tapfds[pair_idx] = -1;
+ tapfd = dev->tapfds[pair_idx];
+ if (vhost_kernel_set_backend(vhostfd, -1) < 0) {
+ PMD_DRV_LOG(ERR, "fail to set backend for vhost kernel");
+ return -1;
}
- return vhost_kernel_set_backend(vhostfd, -1);
- } else if (dev->tapfds[pair_idx] >= 0) {
+ if (req_mq && vhost_kernel_tap_set_queue(tapfd, false) < 0) {
+ PMD_DRV_LOG(ERR, "fail to disable tap for vhost kernel");
+ return -1;
+ }
+ dev->qp_enabled[pair_idx] = false;
return 0;
}
+ if (dev->tapfds[pair_idx] >= 0) {
+ tapfd = dev->tapfds[pair_idx];
+ if (vhost_kernel_tap_set_offload(tapfd, dev->features) == -1)
+ return -1;
+ if (req_mq && vhost_kernel_tap_set_queue(tapfd, true) < 0) {
+ PMD_DRV_LOG(ERR, "fail to enable tap for vhost kernel");
+ return -1;
+ }
+ goto set_backend;
+ }
+
if ((dev->features & (1ULL << VIRTIO_NET_F_MRG_RXBUF)) ||
(dev->features & (1ULL << VIRTIO_F_VERSION_1)))
hdr_size = sizeof(struct virtio_net_hdr_mrg_rxbuf);
@@ -353,13 +371,15 @@ vhost_kernel_enable_queue_pair(struct virtio_user_dev *dev,
return -1;
}
+ dev->tapfds[pair_idx] = tapfd;
+
+set_backend:
if (vhost_kernel_set_backend(vhostfd, tapfd) < 0) {
PMD_DRV_LOG(ERR, "fail to set backend for vhost kernel");
- close(tapfd);
return -1;
}
- dev->tapfds[pair_idx] = tapfd;
+ dev->qp_enabled[pair_idx] = true;
return 0;
}
@@ -18,7 +18,7 @@
#include "../virtio_logs.h"
#include "../virtio_pci.h"
-static int
+int
vhost_kernel_tap_set_offload(int fd, uint64_t features)
{
unsigned int offload = 0;
@@ -37,26 +37,34 @@ vhost_kernel_tap_set_offload(int fd, uint64_t features)
offload |= TUN_F_UFO;
}
- if (offload != 0) {
- /* Check if our kernel supports TUNSETOFFLOAD */
- if (ioctl(fd, TUNSETOFFLOAD, 0) != 0 && errno == EINVAL) {
- PMD_DRV_LOG(ERR, "Kernel does't support TUNSETOFFLOAD\n");
- return -ENOTSUP;
- }
+ /* Check if our kernel supports TUNSETOFFLOAD */
+ if (ioctl(fd, TUNSETOFFLOAD, 0) != 0 && errno == EINVAL) {
+ PMD_DRV_LOG(ERR, "Kernel does't support TUNSETOFFLOAD\n");
+ return -ENOTSUP;
+ }
+ if (ioctl(fd, TUNSETOFFLOAD, offload) != 0) {
+ offload &= ~TUN_F_UFO;
if (ioctl(fd, TUNSETOFFLOAD, offload) != 0) {
- offload &= ~TUN_F_UFO;
- if (ioctl(fd, TUNSETOFFLOAD, offload) != 0) {
- PMD_DRV_LOG(ERR, "TUNSETOFFLOAD ioctl() failed: %s\n",
- strerror(errno));
- return -1;
- }
+ PMD_DRV_LOG(ERR, "TUNSETOFFLOAD ioctl() failed: %s\n",
+ strerror(errno));
+ return -1;
}
}
return 0;
}
+int
+vhost_kernel_tap_set_queue(int fd, bool attach)
+{
+ struct ifreq ifr = {
+ .ifr_flags = attach ? IFF_ATTACH_QUEUE : IFF_DETACH_QUEUE,
+ };
+
+ return ioctl(fd, TUNSETQUEUE, &ifr);
+}
+
int
vhost_kernel_open_tap(char **p_ifname, int hdr_size, int req_mq,
const char *mac, uint64_t features)
@@ -2,6 +2,10 @@
* Copyright(c) 2016 Intel Corporation
*/
+#ifndef _VHOST_KERNEL_TAP_H
+#define _VHOST_KERNEL_TAP_H
+
+#include <stdbool.h>
#include <sys/ioctl.h>
/* TUN ioctls */
@@ -37,3 +41,7 @@
int vhost_kernel_open_tap(char **p_ifname, int hdr_size, int req_mq,
const char *mac, uint64_t features);
+int vhost_kernel_tap_set_offload(int fd, uint64_t features);
+int vhost_kernel_tap_set_queue(int fd, bool attach);
+
+#endif
@@ -456,6 +456,9 @@ vhost_user_enable_queue_pair(struct virtio_user_dev *dev,
{
int i;
+ if (dev->qp_enabled[pair_idx] == enable)
+ return 0;
+
for (i = 0; i < 2; ++i) {
struct vhost_vring_state state = {
.index = pair_idx * 2 + i,
@@ -466,6 +469,7 @@ vhost_user_enable_queue_pair(struct virtio_user_dev *dev,
return -1;
}
+ dev->qp_enabled[pair_idx] = enable;
return 0;
}
@@ -545,8 +545,11 @@ virtio_user_dev_uninit(struct virtio_user_dev *dev)
}
if (dev->vhostfds) {
- for (i = 0; i < dev->max_queue_pairs; ++i)
+ for (i = 0; i < dev->max_queue_pairs; ++i) {
close(dev->vhostfds[i]);
+ if (dev->tapfds[i] >= 0)
+ close(dev->tapfds[i]);
+ }
free(dev->vhostfds);
free(dev->tapfds);
}
@@ -49,6 +49,7 @@ struct virtio_user_dev {
struct vring_packed packed_vrings[VIRTIO_MAX_VIRTQUEUES];
};
struct virtio_user_queue packed_queues[VIRTIO_MAX_VIRTQUEUES];
+ bool qp_enabled[VIRTIO_MAX_VIRTQUEUE_PAIRS];
struct virtio_user_backend_ops *ops;
pthread_mutex_t mutex;