Re: [PATCH repost 3] [SCSI] Retrieve the Caching mode page

[James Bottomley]

On Wed, 2010-11-24 at 09:49 -0500, Alan Stern wrote:
> On Wed, 24 Nov 2010, James Bottomley wrote:
> 
> > On Wed, 2010-11-24 at 01:02 -0800, Luben Tuikov wrote:
> > > I doubt this as very unlikely. Has anyone actually seen a device that
> > > sends mode parameter data with faux Caching mode page or corrupted
> > > data that is in fact interpreted as a Caching mode page? Is such a
> > > device fully operational sans the faux Caching mode page, or does it
> > > just not work? Is it common to have devices having a faux Caching mode
> > > page or corrupted mode parameter data resulting in a Caching mode page
> > > with random data?
> > > 
> > > Undoubtedly, as the usb-storage maintainer, you must have variety of
> > > devices, some broken some not. Could you apply this patch to your tree
> > > and test some of the devices you have? My tests indicate a stable
> > > behavior.
> > 
> > The basic problem isn't devices lying ... the worst we'll do is current
> > behaviour (not SYNC when we should).  The problem is devices that get
> > confused (or worse simply crash the firmware).  The best way to avoid
> > the crashing firmware problem ... if we can assume that modern USB
> > devices are better is to key off the SCSI version.  Unfortunately, in
> > spite of several attempts, we've never managed to stop usbstorage lying
> > about this:
> > 
> > 		/* Some devices report a SCSI revision level above 2 but are
> > 		 * unable to handle the REPORT LUNS command (for which
> > 		 * support is mandatory at level 3).  Since we already have
> > 		 * a Get-Max-LUN request, we won't lose much by setting the
> > 		 * revision level down to 2.  The only devices that would be
> > 		 * affected are those with sparse LUNs. */
> > 		if (sdev->scsi_level > SCSI_2)
> > 			sdev->sdev_target->scsi_level =
> > 					sdev->scsi_level = SCSI_2;
> > 
> > Untangling all of this would be rather complex, I fear.
> 
> Quite likely.
> 
> > The final question is is it worth it?  Since USB devices are supposed to
> > be hot unpluggable, surely a USB device with a write back cache would be
> > a disaster: no-one will SYNC the cache on a surprise unplug anyway ...
> > therefore there shouldn't really be any of them surviving in the wild
> > (famous last words, I suppose).
> 
> Well, hot unpluggable doesn't mean it's okay to unplug the device at 
> any time.  For example, under Windows you're not supposed to unplug a 
> USB drive without first going through the "Safely remove hardware" 
> applet.  And of course, you can easily guess what command that applet 
> sends to the device...
> 
> On the whole, I'm with Luben on this.  The likelihood of introducing
> bad behavior because of devices sending incorrect cache-page
> information seems very small.

Yes, just assure me that sending the mode page request won't kill
anything and I'm fine with applying it.  As I said, as long as we get a
reply, whatever it is it can't make use behave any worse than we do now.

James


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/