Re: 2.6.26-rc9: Reported regressions from 2.6.25

[James Bottomley <James.Bottomley-d9PhHud1JfjCXq6kfMZ53/egYHeGw8Jk@public.gmane.org>]

On Sun, 2008-07-06 at 08:46 -0700, Linus Torvalds wrote:
> > Bug-Entry     : http://bugzilla.kernel.org/show_bug.cgi?id=10725
> > Subject               : USB Mass storage mount fails: Write protect on
> > Submitter     : Maciej Rutecki <maciej.rutecki-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> > Date          : 2008-05-16 14:55 (52 days old)
> > References    : http://marc.info/?l=linux-kernel&m=121095168003572&w=4
> > Handled-By    : Alan Stern <stern-nwvwT67g6+6dFdvTe/nMLpVzexx5G7lz@public.gmane.org>
> > Patch         : http://marc.info/?l=linux-scsi&m=121433068314568&w=2
> 
> Hmm. James?

Oh ... it was reported against 2.6.26-rc1, so the fix is actually
crrently queued for scsi-misc in my internal test rig.  However, the
reason for the long QA is that there's a potential danger to the fix
which is clearing the buffer from the reported length to the actual
length will break a SCSI card that misreports the returned length.  It
seems to be OK on my esoteric SCSI card collection, so I'll push it out
to scsi-rc-fixes.  However, I think it still wants to run in linux-next
for a few days to see if anyone else can turn up a problem.

The true issue, of course, is that we won't see a problem until the
2.6.26 release because the hardware that triggers it isn't in the set
we're testing with.  A less risky fix for 2.6.26-rc9 might be to move
the clearing into the affected subsystem (usbstorage).

James