HI Jiayu, On 3/28/22 09:04, David Marchand wrote:
On Mon, Mar 28, 2022 at 4:08 AM Jiayu Hu <jiayu...@intel.com> wrote:NULL check for vq->async must be protected by lock. Otherwise, it is possible that the data plane thread dereferences vq->async with NULL value, since the control plane thread is freeing vq->async. Fixes: ee8024b3d4ad (vhost: move async data in dedicated structure) Cc: sta...@dpdk.org Signed-off-by: Jiayu Hu <jiayu...@intel.com> --- lib/vhost/vhost.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/lib/vhost/vhost.c b/lib/vhost/vhost.c index bc88148347..7f60c2824f 100644 --- a/lib/vhost/vhost.c +++ b/lib/vhost/vhost.c @@ -1887,9 +1887,6 @@ rte_vhost_async_get_inflight(int vid, uint16_t queue_id) if (vq == NULL) return ret; - if (!vq->async) - return ret; - if (!rte_spinlock_trylock(&vq->access_lock)) { VHOST_LOG_CONFIG(DEBUG, "(%s) failed to check in-flight packets. virtqueue busy.\n", @@ -1897,6 +1894,9 @@ rte_vhost_async_get_inflight(int vid, uint16_t queue_id) return ret; } + if (!vq->async) + return ret;Lock is still taken at this point. FYI, I'll post a series to instrument locks in vhost, soon.
Could you please send a v2 which does not return with the lock taken?
+ ret = vq->async->pkts_inflight_n; rte_spinlock_unlock(&vq->access_lock);
Thanks, Maxime
