[PATCH 04/24] scsi-multipath: introduce scsi_mpath_device_class

John Garry john.g.garry at oracle.com
Tue Mar 3 02:53:35 PST 2026


On 03/03/2026 07:16, Hannes Reinecke wrote:
> On 2/25/26 16:36, John Garry wrote:
>> Introduce a new class for multipathed devices, scsi_mpath_device_class.
>>
>> The purpose of this class is for managing the scsi_mpath_head.dev member.
>>
>> The naming for the scsi_device structure is in form H:C:I:L,
>> where H is host, C is channel, I is ID, and L is lun.
>>
>> However, for a multipathed scsi_device, all the naming members may be
>> different between member scsi_device's. As such, just use a simple
>> single-number naming index for each scsi_mpath_head.
>>
>> The sysfs device folder will have links to the scsi_device's so, it will
>> be possible to lookup the member scsi_device's.
>>
>> An example sysfs entry is as follows:
>> # ls -l /sys/class/scsi_mpath_device/0/
>> total 0
>> drwxr-xr-x    2 root     root             0 Feb 24 11:56 power
>> lrwxrwxrwx    1 root     root             0 Feb 24 11:56 subsystem - 
>> > ../../../../class/scsi_mpath_device
>> -rw-r--r--    1 root     root          4096 Feb 24 11:55 uevent
>> -r--r--r--    1 root     root          4096 Feb 24 11:56 wwid
>> # cat /sys/class/scsi_mpath_device/0/wwid
>> naa.600140505200a986f0043c9afa1fd077
>>
> Ah, here it is.
> So you can ignore my comments from the previous patch.
> (and you might think of merging this and the previous patch).

ok, I can reorder and/or merge.

> 
> But device naming is still dodgy. A plain number has so many ways of
> being misinterpreted.
> Wouldn't it be better to name it 'mpathX' ?
> 

sure, maybe that is better. Any other naming suggestions? Originally I 
had smpdX (but dropped it), where that is "Scsi MultiPath Device" and X 
is a number index.

Cheers



More information about the Linux-nvme mailing list