Re: [PATCH 6/6] scsi: hisi_sas: Expose multiple hw queues for v3 as experimental

[John Garry <john.garry@huawei.com>]

On 27/10/2019 08:19, Ming Lei wrote:
>>   	.this_id		= -1,
>> @@ -3265,8 +3300,14 @@ hisi_sas_v3_probe(struct pci_dev *pdev, const struct pci_device_id *id)
>>   	shost->max_lun = ~0;
>>   	shost->max_channel = 1;
>>   	shost->max_cmd_len = 16;
>> -	shost->can_queue = HISI_SAS_UNRESERVED_IPTT;
>> -	shost->cmd_per_lun = HISI_SAS_UNRESERVED_IPTT;
>> +

Hi Ming,

I mentioned in the thread "blk-mq: improvement on handling IO during CPU 
hotplug" that I was using this series to test that patchset.

So just with this patchset (and without yours), I get what looks like 
some IO errors in the LLDD. The error is an underflow error. I can't 
figure out what is the cause.

I'm wondering if the SCSI command is getting corrupted someway.

>> +	if (expose_mq_experimental) {
>> +		shost->can_queue = HISI_SAS_MAX_COMMANDS;
>> +		shost->cmd_per_lun = HISI_SAS_MAX_COMMANDS;
> The above is contradictory with current 'nr_hw_queues''s meaning,
> see commit on Scsi_Host.nr_hw_queues.
> 

Right, so I am generating the hostwide tag in the LLDD. And the Scsi 
host-wide host_busy counter should ensure that we don't pump too much IO 
to the HBA.

What other problem will this cause?

Thanks,
John

> 
>          /*
>           * In scsi-mq mode, the number of hardware queues supported by the LLD.
>           *
>           * Note: it is assumed that each hardware queue has a queue depth of
>           * can_queue. In other words, the total queue depth per host
>           * is nr_hw_queues * can_queue.
>           */
> 
> Also this implementation wastes memory too much.
> 
> 
> thanks,
> Ming