Re: [PATCH for-next] RDMA/hns: Clear extended doorbell info before using

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 2021/6/9 21:44, Leon Romanovsky wrote:
> On Wed, Jun 09, 2021 at 07:11:05PM +0800, Weihang Li wrote:
>> From: Xi Wang <wangxi11@xxxxxxxxxx>
>>
>> Both of HIP08 and HIP09 require the extended doorbell information to be
>> cleared before being used.
> 
> Is it bugfix or feature?
> For the fix, it needs to have Fixes ... line.
> 

It is a bugfix, I will add a fixes tag, thanks.

>>
>> Signed-off-by: Xi Wang <wangxi11@xxxxxxxxxx>
>> Signed-off-by: Weihang Li <liweihang@xxxxxxxxxx>
>> ---
>>  drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 16 ++++++++++++++++
>>  drivers/infiniband/hw/hns/hns_roce_hw_v2.h |  1 +
>>  2 files changed, 17 insertions(+)
>>
>> diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
>> index fbc45b9..c5d2cfb 100644
>> --- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
>> +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
>> @@ -1572,6 +1572,20 @@ static void hns_roce_function_clear(struct hns_roce_dev *hr_dev)
>>  	}
>>  }
>>  
>> +static void hns_roce_clear_extdb_list_info(struct hns_roce_dev *hr_dev)
>> +{
>> +	struct hns_roce_cmq_desc desc;
>> +	int ret;
>> +
>> +	hns_roce_cmq_setup_basic_desc(&desc, HNS_ROCE_OPC_CLEAR_EXTDB_LIST_INFO,
>> +				      false);
>> +	ret = hns_roce_cmq_send(hr_dev, &desc, 1);
>> +	if (ret)
>> +		ibdev_warn(&hr_dev->ib_dev,
>> +			   "failed to clear extended doorbell info, ret = %d.\n",
>> +			   ret);
>> +}
>> +
>>  static int hns_roce_query_fw_ver(struct hns_roce_dev *hr_dev)
>>  {
>>  	struct hns_roce_query_fw_info *resp;
>> @@ -2684,6 +2698,8 @@ static int hns_roce_v2_init(struct hns_roce_dev *hr_dev)
>>  	if (ret)
>>  		return ret;
>>  
>> +	/* The hns ROCEE requires the extdb info to be cleared before using */
>> +	hns_roce_clear_extdb_list_info(hr_dev);
> 
> If it "requires", why do you proceed anyway? Why don't you check for
> success/failure?
> 
> Thanks
> 

You are right, we should check for it's return value.

Weihang

>>  	if (hr_dev->is_vf)
>>  		return 0;
>>  
>> diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.h b/drivers/infiniband/hw/hns/hns_roce_hw_v2.h
>> index cd361c0..073e835 100644
>> --- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.h
>> +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.h
>> @@ -250,6 +250,7 @@ enum hns_roce_opcode_type {
>>  	HNS_ROCE_OPC_CLR_SCCC				= 0x8509,
>>  	HNS_ROCE_OPC_QUERY_SCCC				= 0x850a,
>>  	HNS_ROCE_OPC_RESET_SCCC				= 0x850b,
>> +	HNS_ROCE_OPC_CLEAR_EXTDB_LIST_INFO		= 0x850d,
>>  	HNS_ROCE_OPC_QUERY_VF_RES			= 0x850e,
>>  	HNS_ROCE_OPC_CFG_GMV_TBL			= 0x850f,
>>  	HNS_ROCE_OPC_CFG_GMV_BT				= 0x8510,
>> -- 
>> 2.7.4
>>
> 




[Index of Archives]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Photo]     [Yosemite News]     [Yosemite Photos]     [Linux Kernel]     [Linux SCSI]     [XFree86]

  Powered by Linux