Re: [PATCH] add_partition silently ignored errors

[Dirk Hohndel <hohndel@linux.intel.com>]

On Fri, Nov 02, 2007 at 03:50:29PM -0400, Bob Copeland wrote:
> On 11/2/07, Dirk Hohndel <hohndel@linux.intel.com> wrote:
> > > > @@ -554,8 +573,11 @@ int rescan_partitions(struct gendisk *disk, struct block_device *bdev)
> > > >             if (from + size > get_capacity(disk)) {
> > > >                     printk(" %s: p%d exceeds device capacity\n",
> > > >                             disk->disk_name, p);
> > > > +                   return -EBUSY;
> [snip]
> > I was wondering about that myself - EBUSY seemed to be used in a couple of
> > other cases where there wasn't a clear match, but I think EOVERFLOW actually
> > might make more sense. Opinions?
> 
> ISTR that some people wanted to keep going in this case rather than
> return an error, e.g. for forensic purposes...
> 
> .. digging... here's a thread from last year:
> 
> http://lkml.org/lkml/2006/5/11/64

Thanks for finding that! I took a different approach than Andries but can
appreciate the argument. I'll remove that line from my patch.

/D