[PATCH 6/7] nvme/nvme-fabrics: introduce nvmf_reconnect_ctrl_work API

Max Gurtovoy mgurtovoy at nvidia.com
Tue Oct 19 06:18:49 PDT 2021


On 10/19/2021 3:44 PM, Sagi Grimberg wrote:
>
>>   static void nvme_tcp_teardown_ctrl(struct nvme_ctrl *ctrl, bool 
>> shutdown)
>> @@ -2435,6 +2416,7 @@ static const struct nvme_ctrl_ops 
>> nvme_tcp_ctrl_ops = {
>>       .get_address            = nvmf_get_address,
>>       .teardown_ctrl_io_queues    = _nvme_tcp_teardown_io_queues,
>>       .teardown_ctrl_admin_queue    = _nvme_tcp_teardown_admin_queue,
>> +    .setup_ctrl            = _nvme_tcp_setup_ctrl,
>
> This asymmetry is not great, what made you choose to go this way?

There are more duplicated flows we can work on in the future.




More information about the Linux-nvme mailing list