Re: [PATCH] pull scsi_scan_host out of scsi_add_host

[Mike Anderson <andmike@us.ibm.com>]

Oliver Neukum [oliver@neukum.org] wrote:
> Am Dienstag, 8. Juli 2003 10:03 schrieb Mike Anderson:
> > Christoph Hellwig [hch@lst.de] wrote:
> > > As Douglas mentioned on this list a while ago there's hardware these
> > > days where we don't want to automatically perform an initial bus scan
> > > in scsi_add_host.  Currently we only have ieee1394 in the tree but
> > > I can imagine more uses like iscsi or certain fc setups.
> > > 
> > > This patch removes the call to scsi_scan_host from scsi_add_host
> > > and moves it into the drivers that want in.
> > 
> > Would we want to make a symmetric interface and pull the
> > unscan out of scsi_remove_host also?
> 
> Unless you are talking about multipath, you have to forget
> all devices connected to a controller, so what sense would
> it make?

Probably not much sense. I was thinking of a progression to user assisted
scanning. In that a driver could never call scsi_scan_host directly and
conversely never call scsi_forget_host directly. 

But in the removal case there is not much response a user space hotplug
handler could make besides a delayed call to scsi_remove_device.

> 
> Philosophically speaking, scanning is an operation that cannot
> be simply undone. You can undo device/driver bindings but you
> cannot undo a bus scan.

Maybe it was a confusing choice of words. The operation I was describing
was to undo the binding.

-andmike
--
Michael Anderson
andmike@us.ibm.com