[PATCH 2/3] nvme: Sync queues on controller resets

[jianchao.w.wang@oracle.com (jianchao.wang)]

Hi Keith

On 02/10/2018 09:55 AM, jianchao.wang wrote:
> Hi Keith
> 
> Thanks for your time and patch for this.
> 
> However, as I shared last time, there is defect here. Please refer to.
> 
> There could be a circular pattern here. Please consider the following scenario:
> 
> timeout_work context                    reset_work context
> nvme_timeout                            nvme_reset_work
>   -> nvme_dev_disable                     -> nvme_sync_queues // hold namespace_mutex
>     -> nvme_stop_queues                     -> blk_sync_queue
>       -> require namespaces_mutex               -> cancel_work_sync(&q->timeout_work)
> 

Looks like we could use rwsem to replace namespaces_mutex.

Thanks
Jianchao


> On the other hand, the blk_mq_kick_requeue_list() should be also added in nvme_kill_queues
> for the case of queue_count < 2
> 
> Thanks
> Jianchao
> 
> On 02/10/2018 01:41 AM, Keith Busch wrote:
>> This patch has the nvme pci driver synchronize request queues to ensure
>> starting the controller is not racing with a previously running timeout
>> handler.
>>
>> Reported-by: Jianchao Wang <jianchao.w.wang at oracle.com>
>> Signed-off-by: Keith Busch <keith.busch at intel.com>
>> ---
>>  drivers/nvme/host/core.c | 15 ++++++++++++++-
>>  drivers/nvme/host/nvme.h |  1 +
>>  drivers/nvme/host/pci.c  |  1 +
>>  3 files changed, 16 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
>> index 2fd8688cfa47..a9bce23a991f 100644
>> --- a/drivers/nvme/host/core.c
>> +++ b/drivers/nvme/host/core.c
>> @@ -3542,12 +3542,25 @@ void nvme_start_queues(struct nvme_ctrl *ctrl)
>>  	struct nvme_ns *ns;
>>  
>>  	mutex_lock(&ctrl->namespaces_mutex);
>> -	list_for_each_entry(ns, &ctrl->namespaces, list)
>> +	list_for_each_entry(ns, &ctrl->namespaces, list) {
>>  		blk_mq_unquiesce_queue(ns->queue);
>> +		blk_mq_kick_requeue_list(ns->queue);
>> +	}
>>  	mutex_unlock(&ctrl->namespaces_mutex);
>>  }
>>  EXPORT_SYMBOL_GPL(nvme_start_queues);
>>  
>> +void nvme_sync_queues(struct nvme_ctrl *ctrl)
>> +{
>> +	struct nvme_ns *ns;
>> +
>> +	mutex_lock(&ctrl->namespaces_mutex);
>> +	list_for_each_entry(ns, &ctrl->namespaces, list)
>> +		blk_sync_queue(ns->queue);
>> +	mutex_unlock(&ctrl->namespaces_mutex);
>> +}
>> +EXPORT_SYMBOL_GPL(nvme_sync_queues);
>> +
>>  int nvme_reinit_tagset(struct nvme_ctrl *ctrl, struct blk_mq_tag_set *set)
>>  {
>>  	if (!ctrl->ops->reinit_request)
>> diff --git a/drivers/nvme/host/nvme.h b/drivers/nvme/host/nvme.h
>> index 27e31c00b306..466081e5f680 100644
>> --- a/drivers/nvme/host/nvme.h
>> +++ b/drivers/nvme/host/nvme.h
>> @@ -374,6 +374,7 @@ void nvme_complete_async_event(struct nvme_ctrl *ctrl, __le16 status,
>>  
>>  void nvme_stop_queues(struct nvme_ctrl *ctrl);
>>  void nvme_start_queues(struct nvme_ctrl *ctrl);
>> +void nvme_sync_queues(struct nvme_ctrl *ctrl);
>>  void nvme_kill_queues(struct nvme_ctrl *ctrl);
>>  void nvme_unfreeze(struct nvme_ctrl *ctrl);
>>  void nvme_wait_freeze(struct nvme_ctrl *ctrl);
>> diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c
>> index 90e276c05f79..7a2e4383c468 100644
>> --- a/drivers/nvme/host/pci.c
>> +++ b/drivers/nvme/host/pci.c
>> @@ -2290,6 +2290,7 @@ static void nvme_reset_work(struct work_struct *work)
>>  	 */
>>  	if (dev->ctrl.ctrl_config & NVME_CC_ENABLE)
>>  		nvme_dev_disable(dev, false);
>> +	nvme_sync_queues(&dev->ctrl);
>>  
>>  	/*
>>  	 * Introduce CONNECTING state from nvme-fc/rdma transports to mark the
>>
> 
> _______________________________________________
> Linux-nvme mailing list
> Linux-nvme at lists.infradead.org
> https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.infradead.org_mailman_listinfo_linux-2Dnvme&d=DwICAg&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=7WdAxUBeiTUTCy8v-7zXyr4qk7sx26ATvfo6QSTvZyQ&m=-ffcRNVB1ERVD0CZrnJUyZOFBETbuy0WHX7ksyM8Dcs&s=KWBywqLDbcong1rA6oK0dD_hff4FSr7Y86aUkY-gskE&e=
>