Re: [PATCH v5 03/15] scsi: ufs: implement scsi host timeout handler

[ygardi]

> On 03/03/2016 05:10 PM, ygardi@xxxxxxxxxxxxxx wrote:
>>> On 03/01/2016 09:25 PM, ygardi@xxxxxxxxxxxxxx wrote:
>>>>> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>>>>>> A race condition exists between request requeueing and scsi layer
>>>>>> error handling:
>>>>>> When UFS driver queuecommand returns a busy status for a request,
>>>>>> it will be requeued and its tag will be freed and set to -1.
>>>>>> At the same time it is possible that the request will timeout and
>>>>>> scsi layer will start error handling for it. The scsi layer reuses
>>>>>> the request and its tag to send error related commands to the
>>>>>> device,
>>>>>> however its tag is no longer valid.
>>>>> Hmm. How can the host return a 'busy' status for a request?
>>>>> From my understanding we have three possibilities:
>>>>>
>>>>> 1) queuecommand returns busy; however, that means that the command
>>>>> has
>>>>> never been send and this issue shouldn't occur
>>>>> 2) The command returns with BUSY status. But in this case it has
>>>>> already
>>>>> been returned, so there cannot be any timeout coming in.
>>>>> 3) The host receives a command with a tag which is already in-use.
>>>>> However, that should have been prevented by the block-layer, which
>>>>> really should ensure that this situation never happens.
>>>>>
>>>>> So either way I look at it, it really looks like a bug and adding a
>>>>> timeout handler will just paper over it.
>>>>> (Not that a timeout handler is a bad idea, in fact I'm convinced that
>>>>> you need one. Just not for this purpose.)
>>>>>
>>>>> So can you elaborate how this 'busy' status comes about?
>>>>> Is the command sent to the device?
>>>>>
>>>>> Cheers,
>>>>>
>>>>> Hannes
>>>>
>>>>
>>>> Hi Hannes,
>>>>
>>>> it's going to be a bit long :)
>>>> I think you are missing the point.
>>>> I will describe a race condition happened to us a while ago, that was
>>>> quite difficult to understand and fix.
>>>> So, this patch is not about the "busy" returning to the scsi dispatch
>>>> routine. it's about the abort triggered after 30 seconds.
>>>>
>>>> imagine a request being queued and sent to the scsi, and then to the
>>>> ufs.
>>>> a timer, initialized to 30 seconds start ticking.
>>>> but the request is never sent to the ufs device, as queuecommand()
>>>> returns
>>>> with "SCSI_MLQUEUE_HOST_BUSY"
>>>> by looking at the code, this could happen, for example:
>>>> 	err = ufshcd_hold(hba, true);
>>>> 	if (err) {
>>>> 		err = SCSI_MLQUEUE_HOST_BUSY;
>>>> 		goto out;
>>>> 	}
>>>>
>>> Uuhhh.
>>> You probably should not have pointed me to that piece of code ...
>>> open-coding loops in ufshcd_hold() ... shudder.
>>> (Did I ever review that one? Must've ...)
>>> _Anyway_: sleeping in queuecommand is always a bad idea, as then
>>> precisely those issues you've just described will happen.
>>>
>>> Couldn't you just call
>>> ufshcd_hold(hba, false)
>>> instead of
>>> ufshcd_hold(hba, true)
>>> ?
>>> The request will be requeued more-or-less immediately, avoiding the
>>> issue with timeout handler kicking in.
>>> And the queue will remain blocked until the ungate work item returns,
>>> at
>>> which point I/O submission will continue.
>>> As the request will be requeued to the head of the queue there won't be
>>> other I/O competing with tags, so it shouldn't have any adverse
>>> effects.
>>>
>>> Wouldn't that work?
>>>
>>> Cheers,
>>>
>>> Hannes
>>
>> Hi Hannes
>>
>> This is a bug, and it should be fixed.
> Oh, definitely agreed. The question is _where_.
>
>
>> if you choose to bypass it, by calling ufshcd_hold(hba, false), not only
>> the race condition is still there, and can pop-out at any other point in
>> the future, but also, not sure what are the consequences of
>> ufshcd_hold(hba, false) unstead of "true".
> Well ... seeing it's your driver, I would've thought _you_ should know ...
>
>> so, changing the already tested and working code, (not to return BUSY
>> from
>> queuecommand) is not a fix.
> Hey, I did _not_ suggest not to retury BUSY from queuecommand.
>
> I was suggesting this patch:
>
> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
> index 9c1b94b..b9295ad 100644
> --- a/drivers/scsi/ufs/ufshcd.c
> +++ b/drivers/scsi/ufs/ufshcd.c
> @@ -1388,7 +1388,7 @@ static int ufshcd_queuecommand(struct Scsi_Host
> *host, struct scsi_cmnd *cmd)
>                 goto out;
>         }
>
> -       err = ufshcd_hold(hba, true);
> +       err = ufshcd_hold(hba, false);
>         if (err) {
>                 err = SCSI_MLQUEUE_HOST_BUSY;
>                 clear_bit_unlock(tag, &hba->lrb_in_use);
>
> which, by reading the code, should be avoiding this issue.


Hannes,
we are not trying to avoid returning BUSY from queuecommand().
On the contrary. By returning BUSY we actually re-queuing the request
which is exactly what we need to do.
your patch doesn't fix the race condition.

thanks,
Yaniv

> I was just asking you if you could give this patch a spin and see if it
> works. If not (for whatever reason) I'm happy to accept your patch.
> But first I would like to have an explanation why the above would _not_
> work.
>
> Unfortunately I don't have the hardware otherwise I'd be running the
> tests myself.
>
> Cheers,
>
> Hannes
> --
> Dr. Hannes Reinecke		      zSeries & Storage
> hare@xxxxxxx			      +49 911 74053 688
> SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
> GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
> --
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>