Re: [PATCH 2.6.13 5/14] sas-class: sas_discover.c Discover process (end devices)

[Patrick Mansfield <patmans@us.ibm.com>]

On Wed, Sep 14, 2005 at 08:28:29AM -0400, Luben Tuikov wrote:
> On 09/13/05 18:42, Patrick Mansfield wrote:

> > What I mean is that the target has to intercept the command whether it is
> > a REPORT LUN or for the well known (W_LUN).
> > 
> > The target (firmware) code has to have code today like:
> > 
> > 	if (cmd == REPORT_LUN) {
> > 		do_report_lun();
> > 	}
> > 
> > For only W_LUN support, the code might be something like:
> > 
> > 	if (lun == W_LUN) {
> > 		if (cmd == REPORT_LUN) {
> > 			do_report_lun();
> > 		}
> > 	}
> > 
> > But the first case above already covers even the W_LUN case.
> 
> _Except_, that what the firmware actually does is, it routes
> the tasks by LUN first, _before_ looking up with what the command
> is.*  This is crucial.

That is what the second code snippet above is meant to show.

> > So adding a W_LUN at this point does not add any value ... maybe it looks
> > nice in the spec and in someones firmware, but it does not add anything
> > that I can see.
> 
> I wonder if the maintainer of the SCSI Core would listen or ignore your
> opinion here.

> I wonder _who_ decides here where speculation ends and industry
> opinion starts?

I am talking about the scsi spec, not the code. IMO linux scsi code should
support W_LUN and 64 bit LUN.

-- Patrick Mansfield