Re: [PATCH v3 1/6] [SCSI] Add a kernel config option for enhanced Unit Attention support

[Ewan Milne <emilne@redhat.com>]

On Wed, 2013-06-19 at 18:35 +0000, James Bottomley wrote:
> On Wed, 2013-06-19 at 13:42 -0400, Ewan D. Milne wrote:
> > From: "Ewan D. Milne" <emilne@redhat.com>
> > 
> > Added CONFIG_SCSI_ENHANCED_UA kernel config option to enable changes
> > in the SCSI mid-layer which detect and report certain Unit Attention
> > conditions reported by devices.  These changes are primarily useful
> > when storage arrays that can be reconfigured are being used, so the
> > config option would normally not be used unless it was needed.
> > 
> > Signed-off-by: Ewan D. Milne <emilne@redhat.com>
> > ---
> >  drivers/scsi/Kconfig | 12 ++++++++++++
> >  1 file changed, 12 insertions(+)
> > 
> > diff --git a/drivers/scsi/Kconfig b/drivers/scsi/Kconfig
> > index e955978..5d1e614 100644
> > --- a/drivers/scsi/Kconfig
> > +++ b/drivers/scsi/Kconfig
> > @@ -280,6 +280,18 @@ config SCSI_WAIT_SCAN
> >  # disabling it, whereas people who accidentally switch it off may wonder why
> >  # their mkinitrd gets into trouble.
> >  
> > +config SCSI_ENHANCED_UA
> > +       bool "Enhanced SCSI Unit Attention handling"
> > +       depends on SCSI
> > +       help
> > +	 Certain SCSI devices report changes via a UNIT ATTENTION code.
> > +	 (For example, the addition or removal of LUNs from a target, or
> > +	 changing the number of logical blocks on a LUN.)  This option
> > +	 enables reporting of these changes via udev events, so that the
> > +	 device can be rescanned to find out what has changed.  This is
> > +	 primarily useful when storage arrays that can be reconfigured
> > +	 are attached to the system, otherwise you can say N here.
> > +
> 
> I don't think we need this.  If we're going to do device events, we
> should do it unconditionally.  The recourse for not wanting them is not
> listening for them.  That way we don't have two separate config branches
> to maintain, one of which will surely bitrot.

OK.  I added it because I thought this wouldn't be useful for desktop
and/or embedded kernels, but I'm happy if it is not needed.

-Ewan

> 
> James
>