Re: [PATCH] tcm_vhost: Multi-target support

[Asias He <asias@redhat.com>]

On 02/01/2013 03:38 PM, Nicholas A. Bellinger wrote:
> On Fri, 2013-02-01 at 12:03 +0800, Asias He wrote:
>> On 02/01/2013 04:59 AM, Nicholas A. Bellinger wrote:
>>> On Thu, 2013-01-31 at 17:28 +0800, Asias He wrote:
>>>> Hello Nicholas,
>>>>
>>>> On 01/31/2013 03:33 PM, Asias He wrote:
>>>>> In order to take advantages of Paolo's multi-queue virito-scsi, we need
>>>>> multi-target support in tcm_vhost first. Otherwise all the requests go
>>>>> to one queue and other queues are idle.
>>>>>
> 
> <SNIP>
> 
>>>>> @@ -771,14 +799,11 @@ static int vhost_scsi_set_endpoint(
>>>>>  		}
>>>>>  		tv_tport = tv_tpg->tport;
>>>>>  
>>>>> -		if (!strcmp(tv_tport->tport_name, t->vhost_wwpn) &&
>>>>> -		    (tv_tpg->tport_tpgt == t->vhost_tpgt)) {
>>>>> +		if (!strcmp(tv_tport->tport_name, t->vhost_wwpn)) {
>>>>>  			tv_tpg->tv_tpg_vhost_count++;
>>>>> -			mutex_unlock(&tv_tpg->tv_tpg_mutex);
>>>>> -			mutex_unlock(&tcm_vhost_mutex);
>>>>>  
>>>>>  			mutex_lock(&vs->dev.mutex);
>>>>> -			if (vs->vs_tpg) {
>>>>> +			if (vs->vs_tpg[tv_tpg->tport_tpgt - 1]) {
>>>>>  				mutex_unlock(&vs->dev.mutex);
>>>>>  				mutex_lock(&tv_tpg->tv_tpg_mutex);
>>>>>  				tv_tpg->tv_tpg_vhost_count--;
>>>>> @@ -786,15 +811,17 @@ static int vhost_scsi_set_endpoint(
>>>>>  				return -EEXIST;
>>>>>  			}
>>>>>  
>>>>> -			vs->vs_tpg = tv_tpg;
>>>>> +			vs->vs_tpg[tv_tpg->tport_tpgt - 1] = tv_tpg;
>>>>
>>>>
>>>> tv_tpg->tport_tpgt starts from 0, right? I thought it starts from 1,
>>>> because I always got it starts from 1 in targetcli.
>>>>
>>>> o- vhost
>>>>    o- naa.6001405bd4e8476d
>>>>       o- tpg1
>>>>          o- luns
>>>>             o- lun0
>>>>       o- tpg2
>>>>          o- luns
>>>>             o- lun0
>>>>       o- tpg3
>>>>          o- luns
>>>>             o- lun0
>>>>       o- tpg4
>>>>          o- luns
>>>>             o- lun0
>>>>
>>>
>>> So at least with iscsi-target, we start from tpgt=1 to avoid some legacy
>>> initiators that have issues handling tgpt=0.
>>>
>>> Given that rtslib/targetcli currently expect this with the "tpgs"
>>> feature is enabled, starting from tpgt=1 with tcm_vhost probably makes
>>> the most sense.
>>
>> Okay. But tgpt can be 0, right?
>>
> 
> Most certainly, in the end it's totally up to the fabric.  ;)

okay.

>  
>> I saw this setup:
>>
>>   cd /sys/kernel/config/target
>>   mkdir -p core/fileio_0/fileio
>>   echo 'fd_dev_name=/home/pbonzini/test.img,fd_dev_size=5905580032' >
>> core/fileio_0/fileio/control
>>   echo 1 > core/fileio_0/fileio/enable
>>   mkdir -p vhost/naa.600140554cf3a18e/tpgt_0/lun/lun_0
>>   cd vhost/naa.600140554cf3a18e/tpgt_0
>>   ln -sf ../../../../../core/fileio_0/fileio/ lun/lun_0/virtual_scsi_port
>>   echo naa.60014053226f0388 > nexus
>>
>> And this:
>>
>>        ** Setup wwpn and tpgt
>>        $ wwpn="naa.0"
>>        $ tpgt=/sys/kernel/config/target/vhost/$wwpn/tpgt_0
>>        $ nexus=$tpgt/nexus
>>        $ mkdir -p $tpgt
>>        $ echo -n $wwpn > $nexus
>>
>>
> 
> OK, I think you'll want to avoid the extra vs->vs_tpg[tpgt - 1] offset
> above to properly support this.

Yes. Already did that.

> --nab
> 
>>
>>>> If it is true. I will cook v2 of this patch.
>>>>
>>>> Also, the tv_tpg->tport_tpgt can be none-continuous. e.g.
>>>>
>>>> o- vhost
>>>>    o- naa.6001405bd4e8476d
>>>>       o- tpg1
>>>>          o- luns
>>>>             o- lun0
>>>>       o- tpg2
>>>>          o- luns
>>>>             o- lun0
>>>>       o- tpg4
>>>>          o- luns
>>>>             o- lun0
>>>>
>>>> I will handle this in v2.
>>>>
>>>
>>> Correct, tpgt values may be optionally non-contiguous up to unsigned
>>> short.
>>
>> ok.
>>
>>> --nab
>>>
>>>
>>
>>
> 
> 


-- 
Asias