Re: [PATCH v5 0/7] libsas and drivers: NCQ error handling

[Damien Le Moal <damien.lemoal@opensource.wdc.com>]

On 10/6/22 06:28, Niklas Cassel wrote:
> On Wed, Oct 05, 2022 at 09:53:52AM +0100, John Garry wrote:
>> On 04/10/2022 15:04, John Garry wrote:
>>
>> Hi Niklas,
>>
>> Could you try a change like this on top:
>>
>> void sas_ata_device_link_abort(struct domain_device *device, bool
>> force_reset)
>> {
>> 	struct ata_port *ap = device->sata_dev.ap;
>> 	struct ata_link *link = &ap->link;
>>
>> +	device->sata_dev.fis[2] = ATA_ERR | ATA_DRDY;
>> +	device->sata_dev.fis[3] = 0x04;
>>
>> 	link->eh_info.err_mask |= AC_ERR_DEV;
>> 	if (force_reset)
>> 		link->eh_info.action |= ATA_EH_RESET;
>> 	ata_link_abort(link);
>> }
>> EXPORT_SYMBOL_GPL(sas_ata_device_link_abort);
>>
>> I tried it myself and it looked to work ok, except I have a problem with my
>> arm64 system in that the read log ext times-out and all TF show "device
>> error", like:
> 
> Do you know why it fails to read the log?
> Can you read the NCQ Command Error log using ATA16 passthrough commands?
> 
> sudo sg_sat_read_gplog -d --log=0x10 /dev/sdc
> 
> The first byte is the last NCQ tag (in hex) that failed.

libata issues read log as a non-ncq command under EH. So the NCQ error log
will not help.

> 
> 
> I tried your patch, and it looks good:
> 
> [ 6656.228131] ata5.00: exception Emask 0x0 SAct 0x460000 SErr 0x0 action 0x0
> [ 6656.252759] ata5.00: failed command: WRITE FPDMA QUEUED
> [ 6656.271554] ata5.00: cmd 61/00:00:00:d8:8a/04:00:ce:03:00/40 tag 17 ncq dma 524288 out
>                         res 41/04:00:00:00:00/00:00:00:00:00/00 Emask 0x1 (device error)
> [ 6656.309308] ata5.00: status: { DRDY ERR }
> [ 6656.316403] ata5.00: error: { ABRT }
> [ 6656.322300] ata5.00: failed command: WRITE FPDMA QUEUED
> [ 6656.330871] ata5.00: cmd 61/00:00:00:dc:8a/04:00:ce:03:00/40 tag 18 ncq dma 524288 out
>                         res 41/04:00:00:00:00/00:00:00:00:00/00 Emask 0x1 (device error)
> [ 6656.356295] ata5.00: status: { DRDY ERR }
> [ 6656.362931] ata5.00: error: { ABRT }
> [ 6656.368897] ata5.00: failed command: WRITE FPDMA QUEUED
> [ 6656.377471] ata5.00: cmd 61/00:00:00:d4:8a/04:00:ce:03:00/40 tag 22 ncq dma 524288 out
>                         res 43/04:00:ff:d7:8a/00:00:ce:03:00/40 Emask 0x400 (NCQ error) <F>
> [ 6656.403149] ata5.00: status: { DRDY SENSE ERR }
> [ 6656.410624] ata5.00: error: { ABRT }
> 
> However, since this is a change from the existing behavior of this driver,
> this could go as a separate patch, and does not need to delay this series.
> 
> 
> 
> I also think that we should do a similar patch for sas_ata_task_done():
> 
> diff --git a/drivers/scsi/libsas/sas_ata.c b/drivers/scsi/libsas/sas_ata.c
> index d35c9296f738..648d0693ceee 100644
> --- a/drivers/scsi/libsas/sas_ata.c
> +++ b/drivers/scsi/libsas/sas_ata.c
> @@ -140,7 +140,7 @@ static void sas_ata_task_done(struct sas_task *task)
>                         }
>  
>                         dev->sata_dev.fis[3] = 0x04; /* status err */
> -                       dev->sata_dev.fis[2] = ATA_ERR;
> +                       dev->sata_dev.fis[2] = ATA_ERR | ATA_DRDY;
>                 }
>         }
> 
> To avoid all SAS errors from being reported as HSM errors.
> 
> 
> Kind regards,
> Niklas

-- 
Damien Le Moal
Western Digital Research