On Thu, Feb 16, 2017 at 3:44 PM, Hannes Reinecke <h...@suse.de> wrote:
> On 02/16/2017 11:09 AM, Sreekanth Reddy wrote:
>> On Tue, Jan 31, 2017 at 2:55 PM, Hannes Reinecke <h...@suse.de> wrote:
>>> When sending a TMF via the ioctl interface we should be using
>>> the hi-priority queue instead of the scsi queue to be consistent
>>> with overall TMF usage.
>>>
>>> Signed-off-by: Hannes Reinecke <h...@suse.com>
>>> ---
>>>  drivers/scsi/mpt3sas/mpt3sas_ctl.c | 2 +-
>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/scsi/mpt3sas/mpt3sas_ctl.c 
>>> b/drivers/scsi/mpt3sas/mpt3sas_ctl.c
>>> index 95f0f24..e952175 100644
>>> --- a/drivers/scsi/mpt3sas/mpt3sas_ctl.c
>>> +++ b/drivers/scsi/mpt3sas/mpt3sas_ctl.c
>>> @@ -720,7 +720,7 @@ enum block_state {
>>>                 }
>>>         } else {
>>>
>>> -               smid = mpt3sas_base_get_smid_scsiio(ioc, ioc->ctl_cb_idx, 
>>> NULL);
>>> +               smid = mpt3sas_base_get_smid_hpr(ioc, ioc->ctl_cb_idx);
>>
>> This else part is not for TM path, It is for IO path. For the TM
>> request we are already using smid from hpr queue, as shown below
>>
>> if (mpi_request->Function == MPI2_FUNCTION_SCSI_TASK_MGMT) {
>>                 smid = mpt3sas_base_get_smid_hpr(ioc, ioc->ctl_cb_idx);
>>                 if (!smid) {
>>                         pr_err(MPT3SAS_FMT "%s: failed obtaining a smid\n",
>>                             ioc->name, __func__);
>>                         ret = -EAGAIN;
>>                         goto out;
>>                 }
>>         } else {
>>
>>                 smid = mpt3sas_base_get_smid_scsiio(ioc, ioc->ctl_cb_idx, 
>> NULL);
>>
> Yes, I know.
> But the current method of inserting a SCSI command completely goes
> against blk-mq request usage; with blk-mq the tags are managed with the
> block layer, so you need to integrate with that to get a correct tag.
>
> Is this _really_ a crucial functionality? Can't we just drop it and use
> sg/bsg for this?
> It would make my life _so_ much easier; the alternative would be to
> either implement 'real' reserved command handling or rewriting the above
> code-path in terms of 'struct request', packing and unpacking as we go.
> Neither is very appealing.

I think it is better to take out one request frame from can_queue
count and reserve it for this ioctl scsi io. Since we allow only one
ioctl command at a time.

Thanks,
Sreekanth

>
> Cheers,
>
> Hannes
> --
> Dr. Hannes Reinecke                Teamlead Storage & Networking
> h...@suse.de                                   +49 911 74053 688
> SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
> GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
> HRB 21284 (AG Nürnberg)

Reply via email to