On 04/12/2018 09:35 AM, Maxime Coquelin wrote:

On 04/12/2018 09:34 AM, Chen, Junjie J wrote:

On 04/12/2018 09:21 AM, Tan, Jianfeng wrote:

On 4/12/2018 1:02 AM, Junjie Chen wrote:
dev_start sets *dev_attached* after setup queues, this sets device to
invalid state since no frontend is attached. Also destroy_device set
*started* to zero which makes *allow_queuing* always zero until
dev_start get called again. Actually, we should not determine queues
existence by
*dev_attached* but by queues pointers or other separated variable(s).

Fixes: 30a701a53737 ("net/vhost: fix crash when creating vdev

Signed-off-by: Junjie Chen <junjie.j.c...@intel.com>
Tested-by: Jens Freimann <jfreim...@redhat.com>

Overall, looks great to me except a nit below.

Reviewed-by: Jianfeng Tan <jianfeng....@intel.com>

Thanks Jianfeng, I can handle the small change while applying.

Can you confirm that it is implied that the queue are already allocated, else we wouldn't find the internal resource and quit earlier (in case of eth_dev_close
called twice for example)?

That is required, otherwise it generate segfault if we close device before queue setup. For example we
execute following steps in testpmd:
1. port attach
2. ctrl+D

Thanks for confirming Junjie, I will apply it as is then.

Reviewed-by: Maxime Coquelin <maxime.coque...@redhat.com>

Applied to dpdk-next-virtio/master


Reply via email to