Re: PATCH 2/5: scsi-scan-blist_replun

[Kurt Garloff <garloff@suse.de>]

[-- Attachment #1: Type: text/plain, Size: 2043 bytes --]

Hi Christoph,

On Wed, Apr 21, 2004 at 04:14:12PM +0100, Christoph Hellwig wrote:
> Hmm, where does this number come from?

1 page. I've mentioned it in the email.

> >  	/*
> > -	 * Only support SCSI-3 and up devices.
> > -	 */
> > -	if (sdev->scsi_level < SCSI_3)
> > +	 * Only support SCSI-3 and up devices if BLIST_NOREPORTLUN is not set.
> > +	 * Also allow SCSI-2 if BLIST_REPORTLUN2 is set and host adapter does
> > +	 * support more than 8 LUNs.
> 
> This comment and the nested if is a little confusing.  Also why shouldn't
> the arbitrary SCSI2 limit for BLIST_REPORTLUN2?  I'm sure we won't find
> SCSI1 devices supporting it, but there should not be any harm leaving out
> that check - and doing so makes the logic much cleaner.

It's just straightforward plain simple boolean logic ;-)

I think default_dev_flags=0x20000 would be a good default for many
machines out there, so don't take a risk of breaking SCSI-1 devices.
That's why the check for a HBA which support more than 8 LUNs is in
there as well.

> 	/*
> 	 * REPORT_LUN is used by default only for SCSI 3 devices (unless
> 	 * BLIST_NOREPORTLUN is set).  For other devices BLIST_REPORTLUN
> 	 * enables it if the HBA does support more than 8 LUNs.
> 
> 	 if (bflags & BLIST_NOREPORTLUN)
> 	 	return 1;
> 	 if (sdev->scsi_level < 3 && sdev->host->max_lun <= 8 &&
> 	     !(bflags & BLIST_REPORTLUN2))
> 	     	return 1;
> 
> Note that I'd prefer not to add BLIST_NOREPORTLUN if the discussing
> about the LUN mapping comes to a useable consensus.

I'm sure there will be devices that are SCSI-3 and break on REPORT_LUNS.
It's not just for helping those devs that don't fit into our 32bit LUN
scheme. Also we need to have a replacement for the CONFIG option that
used to be there before.

Regards,
-- 
Kurt Garloff                   <kurt@garloff.de>             [Koeln, DE]
Physics:Plasma modeling <garloff@plasimo.phys.tue.nl> [TU Eindhoven, NL]
Linux: SUSE Labs (Head)        <garloff@suse.de>    [SUSE Nuernberg, DE]

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]