Re: ioccsz and iorcsz check failing

[Max Gurtovoy <mgurtovoy@nvidia.com>]

On 18/12/2023 18:04, Sagi Grimberg wrote:
> 
>>> Is anything else in the identify wrong, or is it just these
>>> fabrics fields?
>>
>> So what I am seeing on the wire are a few fabrics commands (connect, get
>> property) followed by the nvme id ctrl command (opcode 0x6).
>>
>>   nvmet: nvmet_req_init:962
>>   nvmet: nvmet_parse_admin_cmd:1011
>>   nvmet: nvmet_parse_discovery_cmd:359 opcode 6
>>
>> This calls then nvmet_execute_disc_identify, so adding
>>
>> --- a/drivers/nvme/target/discovery.c
>> +++ b/drivers/nvme/target/discovery.c
>> @@ -249,6 +249,7 @@ static void nvmet_execute_disc_identify(struct 
>> nvmet_req *req)
>>   {
>>          struct nvmet_ctrl *ctrl = req->sq->ctrl;
>>          struct nvme_id_ctrl *id;
>> +       u32 cmd_capsule_size;
>>          u16 status = 0;
>>
>>          if (!nvmet_check_transfer_len(req, NVME_IDENTIFY_DATA_SIZE))
>> @@ -289,6 +290,17 @@ static void nvmet_execute_disc_identify(struct 
>> nvmet_req *req)
>>                  id->sgls |= cpu_to_le32(1 << 2);
>>          if (req->port->inline_data_size)
>>                  id->sgls |= cpu_to_le32(1 << 20);
>> +       /*
>> +        * Max command capsule size is sqe + in-capsule data size.
>> +        * Disable in-capsule data for Metadata capable controllers.
>> +        */
>> +       cmd_capsule_size = sizeof(struct nvme_command);
>> +       if (!ctrl->pi_support)
>> +               cmd_capsule_size += req->port->inline_data_size;
>> +       id->ioccsz = cpu_to_le32(cmd_capsule_size / 16);
> 
> Yes, this is the culprit. Nice that it exposed a bug.
> 
> There is no in-capsule data for discovery controllers afaict.

Also, the discovery controllers can't support PI AFAIK.