[PATCH 1/2] nvmet: add get_queue_size op for controllers

Max Gurtovoy mgurtovoy at nvidia.com
Wed Sep 22 00:35:19 PDT 2021


On 9/22/2021 1:47 AM, Sagi Grimberg wrote:
>
>> diff --git a/drivers/nvme/target/core.c b/drivers/nvme/target/core.c
>> index b8425fa34300..cd42655d2980 100644
>> --- a/drivers/nvme/target/core.c
>> +++ b/drivers/nvme/target/core.c
>> @@ -1205,7 +1205,10 @@ static void nvmet_init_cap(struct nvmet_ctrl 
>> *ctrl)
>>       /* CC.EN timeout in 500msec units: */
>>       ctrl->cap |= (15ULL << 24);
>>       /* maximum queue entries supported: */
>> -    ctrl->cap |= NVMET_QUEUE_SIZE - 1;
>> +    if (ctrl->ops->get_queue_size)
>
> Maybe call it max_queue_size? it is a maximum supported in essence.

Yes, I can call it get_max_queue_size.





More information about the Linux-nvme mailing list