On 08/01/2015 04:26 AM, jiang.biao2@xxxxxxxxxx wrote:
> Hannes Reinecke <hare@xxxxxxx> 写于 2015/07/31 21:17:33:
>
>> Hannes Reinecke <hare@xxxxxxx>
>> 2015/07/31 21:17
>>
>> 收件人
>>
>> jiang.biao2@xxxxxxxxxx, linux-scsi@xxxxxxxxxxxxxxx, JBottomley@xxxxxxxx,
>
>>
>> 抄送
>>
>> 主题
>>
>> Re: [Patch] scsi_error: should not get sense for timeout IO in scsi
>> error handler
>>
>> On 07/31/2015 11:52 AM, jiang.biao2@xxxxxxxxxx wrote:
>>> scsi_error: should not get sense for timeout IO in scsi error handler
>>>
>>> When an IO timeout occurs, the IO will be aborted in
>>> scsi_abort_command() and SCSI_EH_ABORT_SCHEDULED will be set. Because
>>> of that, the SCSI_EH_CANCEL_CMD will be clear in scsi_eh_scmd_add().
>>> So when scsi error handler starts, it will get sense for this
>>> timeout IO and the scmd of the IO request will be reused. In that
>>> case, the scmd may be double released when racing with io_done(),
>>> which will result in crash.
>>> SO SCSI_EH_ABORT_SCHEDULED should also be checked when getting sense.
>>> The bug maybe reproduced when the link between host and disk is
>>> unstable.
>>>
>>> Signed-off-by: Jiang Biao <jiang.biao2@xxxxxxxxxx>
>>> Signed-off-by: Long Chun <long.chun@xxxxxxxxxx>
>>> Reviewed-by: Tan Hu <tan.hu@xxxxxxxxxx>
>>> Reviewed-by: Chen Donghai <chen.donghai@xxxxxxxxxx>
>>> Reviewed-by: Cai Qu <cai.qu@xxxxxxxxxx>
>>>
>>> diff -uprN drivers/scsi/scsi_error.c drivers_new/scsi/scsi_error.c
>>> --- scsi/scsi_error.c 2015-07-31 16:03:18.000000000 +0800
>>> +++ scsi_new/scsi_error.c 2015-07-31 16:29:25.000000000 +0800
>>> @@ -1156,9 +1156,14 @@ int scsi_eh_get_sense(struct list_head *
>>> struct Scsi_Host *shost;
>>> int rtn;
>>>
>>> + /*
>>> + * If SCSI_EH_ABORT_SCHEDULED has been set, it is timeout IO,
>>> + * should not get sense.
>>> + */
>>> list_for_each_entry_safe(scmd, next, work_q, eh_entry) {
>>> if ((scmd->eh_eflags & SCSI_EH_CANCEL_CMD) ||
>>> - SCSI_SENSE_VALID(scmd))
>>> + (scmd->eh_eflags & SCSI_EH_ABORT_SCHEDULED) ||
>>> + SCSI_SENSE_VALID(scmd))
>>> continue;
>>>
>>> shost = scmd->device->host;
>>> --
>> _Actually_ you need to test for both, SCSI_EH_CANCEL_CMD _and_
>> SCSI_EH_ABORT_SCHEDULED.
>> Not every driver is required to implement and/or support
>> asynchronous command aborts, and those will be setting
>> SCSI_EH_CANCEL_CMD even though they've run into a timeout.
> That's right, but SCSI_EH_CANCEL_CMD _has_ already been tested
> in the current code, so there's no need to add in the patch.
> After patched, both SCSI_EH_CANCEL_CMD _and_
> SCSI_EH_ABORT_SCHEDULED are tested here, that'll ensure no
> getting sense for timeout io.
>
Ah, right. I've misread the patch.
Reviewed-by: Hannes Reinecke <hare@xxxxxxxx>
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
