On 5/22/23 20:28, John Garry wrote:
> On 22/05/2023 09:00, Jason Yan wrote:
>>
>> OK, so the issue is that __ata_scsi_find_dev() calls ata_find_dev() with
>> devno
>> == scsidev->id. This leads to devno being 0, 1, 2 and 3 for connected
>> drives
>
> This numbering comes from sas_rphy_add():
> ...
> if (identify->device_type == SAS_END_DEVICE &&
> (identify->target_port_protocols &
> (SAS_PROTOCOL_SSP | SAS_PROTOCOL_STP | SAS_PROTOCOL_SATA)))
> rphy->scsi_target_id = sas_host->next_target_id++;
>
> ..
>
> scsi_scan_target(&rphy->dev, 0, rphy->scsi_target_id, lun,
> SCSI_SCAN_INITIAL);
> }
>
> So libata and scsi_transport_sas just use different sdev id numbering
> schemes for host scan.
>
>> sdd, sd1, sdf and sdg, as shown by lsscsi. However, each drive has its own
>> port+link, with the link for each one having ata_link_max_devices() ==
>> 1, so
>> ata_find_dev() works only for the first drive with scsidev->id == 0 and
>> fails
>> for the others. A naive fix would be this:
>>
>> diff --git a/drivers/ata/libata-scsi.c b/drivers/ata/libata-scsi.c
>> index 7bb12deab70c..e4d6f17d7ccc 100644
>> --- a/drivers/ata/libata-scsi.c
>> +++ b/drivers/ata/libata-scsi.c
>> @@ -2718,7 +2718,7 @@ static struct ata_device *__ata_scsi_find_dev(struct
>> ata_port *ap,
>> if (!sata_pmp_attached(ap)) {
>> if (unlikely(scsidev->channel || scsidev->lun))
>> return NULL;
>> - devno = scsidev->id;
>> + devno = 0;
> Would this pattern work:
>
> ata_for_each_dev(ata_dev, link, ALL) {
> if (ata_dev->sdev == sdev)
> return ata_dev;
> }
That would work too I think, even though a loop is a bit ugly...
>
> If not, I think it's ok to have devno = 0 assignment under SAS_HOST
> flag, even though it's far from ideal. Not both of these are not
> preferred, then, as I mentioned before, some per-port callback to do the
> conversion.
See the proper patch I posted a few min ago (I cc-ed you). I do not use SAS_HOST
flag :)
>
> Thanks,
> John
--
Damien Le Moal
Western Digital Research
