On Thu, 2007-04-05 at 08:42 +0100, Christoph Hellwig wrote:
> On Wed, Apr 04, 2007 at 08:05:59PM -0500, James Bottomley wrote:
> > Erm, there's a simple way out of this: That's the BLIST_FORCELUN
> > option. This is why most of the RAID devices have entries in
> > scsi_devinfo.c like:
> >
> > {"ADAPTEC", "AACRAID", NULL, BLIST_FORCELUN},
> >
> > Which overrides the CONFIG_SCSI_MULTI_LUN setting for the particular
> > device. We can easily add an entry for stex as well.
>
> IMHO we should kill CONFIG_SCSI_MULTI_LUN and add blacklist enries for
> devices that don't handle scanning for luns > 1 instead. I think the
> RH/Fedora kernel has started collecting these blacklist entries for
> a long time already.
I wouldn't disagree with that. There have been several schools of
thought on this: One was that every multi-lun device since SPC should
support REPORT LUNS, so MULTI_LUN should be off by default and we
whitelist non SPC compliant multi-lun devices. If we also added a host
hook to allow RAID devices to declare themselves, then I think we
capture 99% of the current cases.
Another is just to make it on by default. I note that both SLES10 and
RHEL5 now define it on, so I can't see much argument in support of a
distro setting it off.
James
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
