Merging se_dev_entry and se_lun?

Merging se_dev_entry and se_lun?

[Hannes Reinecke]

Hi Nic,

having done the patch to export 'write_protect' for demo-mode LUNs
I've came across one puzzling item:

struct se_lun uses a list to refer to the underlying se_dev_entry
structures. Which I found rather curious, as from my understanding
'se_lun' is the structure for the mapped LUN (ie the LUN visible to
the initiator) and 'se_dev_entry' is the underlying physical device
as visible to the LUN.
As such I would have expected a 1:1 relationship between both, ie a
simple pointer from se_lun to se_dev_entry.

Having a list here implies that 'se_lun' can have _several_
se_dev_entry structure attached to it, which I found rather curious.

Can you give me an example where this might be the case?
Or can we replace the list with a simple pointer or even merge both?

Reason I'm asking is the lun_access / dev_flags field; it really
looks like it being a duplicate (I would judge 'write_protect' to be
a property of the mapped LUN, and not of the underlying device),
but in either case having it in both places requires a
synchronisation between both, as for demo-mode LUNs we can only
change it via se_lun, and for others we have to change it via the
se_dev_entry.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		               zSeries & Storage
hare@suse.de			               +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG Nürnberg)

Re: Merging se_dev_entry and se_lun?

[Christoph Hellwig]

On Fri, Jun 19, 2015 at 08:40:23AM +0200, Hannes Reinecke wrote:
> Having a list here implies that 'se_lun' can have _several_
> se_dev_entry structure attached to it, which I found rather curious.
> 
> Can you give me an example where this might be the case?
> Or can we replace the list with a simple pointer or even merge both?

Each initiator has it's own dev entry.

What might make sense with the new list-based dev entry handling is to
merge the se_lun_acl and se_dev_entry, but it would be a fair amount of
work.

Re: Merging se_dev_entry and se_lun?

[Hannes Reinecke]

On 06/19/2015 08:56 AM, Christoph Hellwig wrote:
> On Fri, Jun 19, 2015 at 08:40:23AM +0200, Hannes Reinecke wrote:
>> Having a list here implies that 'se_lun' can have _several_
>> se_dev_entry structure attached to it, which I found rather curious.
>>
>> Can you give me an example where this might be the case?
>> Or can we replace the list with a simple pointer or even merge both?
> 
> Each initiator has it's own dev entry.
> 
But isn't 'se_lun' per initiator, too?

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		               zSeries & Storage
hare@suse.de			               +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG Nürnberg)

Re: Merging se_dev_entry and se_lun?

[Christoph Hellwig]

On Fri, Jun 19, 2015 at 03:21:27PM +0200, Hannes Reinecke wrote:
> > Each initiator has it's own dev entry.
> > 
> But isn't 'se_lun' per initiator, too?

No.
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in