Re: [patch] floppy: suspend/resume fix

[Russell King <rmk+lkml@arm.linux.org.uk>]

On Wed, Nov 15, 2006 at 09:49:33PM +0100, Pavel Machek wrote:
> > > > > Suspending with mounted floppy is a user error.
> > > > 
> > > > Huh?  How so?
> > > 
> > > Floppy is removable, and you are expected to umount removable devices
> > > before suspend.
> > 
> > That seems pretty crude. There are lots of cases where an apparently
> > removable device is/should be preserved properly and left mounted (eg
> > builtin CF).
> > 
> > We really want to be smarter than that - which means the drivers ought to
> > be doing stuff in their suspend/resume paths to figure out if the media
> > changed when really possible (eg IDE removable)
> > 
> > Floppy is probably not too fixable, but calling it a "user error" is
> > insulting - user expectation is reasonable that suspend/resume should
> > just work. The implementation is just rather trickier/nonsensical in this
> > case.
> 
> Yep, it would be nice to do something about that; but I'm not sure how
> this "was media changed" should be implemented, and if it should be
> done in kernel or in userland.

With a floppy the only way to do that is to read data off the disk and
compare it with what we know was on the disk prior to suspend/resume.

Since we don't have a "standard floppy format" (since we're flexible)
you can't rely on MSDOS boot sectors and the like to uniquely identify
floppy disks.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 Serial core