Re: [PATCH] locks: export device name

[Davidlohr Bueso <dave@gnu.org>]

On Tue, 2012-02-14 at 14:09 -0500, J. Bruce Fields wrote:
> On Mon, Feb 13, 2012 at 04:34:25PM -0800, Andrew Morton wrote:
> > On Fri, 10 Feb 2012 22:06:07 +0100
> > Davidlohr Bueso <dave@gnu.org> wrote:
> > 
> > > From: Davidlohr Bueso <dave@gnu.org>
> > > 
> > > The lslk(8) program has not been maintained for over a decade and has recently been rewritten as lslocks(8).
> > > It will be available for the next 2.22 release, in a couple of months. This is a good opportunity to delete
> > > that nasty WE_CAN_BREAK_LSLK_NOW and start exporting the device name instead of the maj:min numbers.
> > > 
> > > For backward compatibility the new version can be in charge of checking older kernel versions and parsing the old
> > > output if necessary.
> > > 
> > > ...
> > >
> > > --- a/fs/locks.c
> > > +++ b/fs/locks.c
> > > @@ -2199,15 +2199,8 @@ static void lock_get_status(struct seq_file *f, struct file_lock *fl,
> > >  			       : (fl->fl_type & F_WRLCK) ? "WRITE" : "READ ");
> > >  	}
> > >  	if (inode) {
> > > -#ifdef WE_CAN_BREAK_LSLK_NOW
> > >  		seq_printf(f, "%d %s:%ld ", fl_pid,
> > >  				inode->i_sb->s_id, inode->i_ino);
> > > -#else
> > > -		/* userspace relies on this representation of dev_t ;-( */
> > > -		seq_printf(f, "%d %02x:%02x:%ld ", fl_pid,
> > > -				MAJOR(inode->i_sb->s_dev),
> > > -				MINOR(inode->i_sb->s_dev), inode->i_ino);
> > > -#endif
> > >  	} else {
> > >  		seq_printf(f, "%d <none>:0 ", fl_pid);
> > >  	}
> > 
> > I don't get it.  This is an immediate and non-back-compatible change to
> > the format of /proc/locks.  The only way this can avoid breaking things
> > is if there are no programs or scripts in use by anyone which use
> > this field.  What am I missing here?
> 
> I'm a little surprised anything parses that file.

To my knowledge only lslk - but the whole point here is that its going
to be replaced by lslocks.

> 
> But, yes, looks like I can "yum install" lslk on Fedora 16, as an
> example.  Can't get it to do anything useful, though.  Does it actually
> work on any recent distro?

It works on Ubuntu's latest release.

> 
> Perhaps safest would be to replace /proc/locks by another interface and
> deprecate this one.

If exporting the name in the current /proc/locks file is out of the
question, then IMHO I don't think it would be worth adding a new
interface just for such a small change.

Thanks,
Davidlohr