Re: [PATCH] add_partition silently ignored errors

[Jens Axboe <jens.axboe@oracle.com>]

On Tue, Oct 30 2007, Dirk Hohndel wrote:
> On Tue, Oct 30, 2007 at 06:31:12PM +0100, Cornelia Huck wrote:
> > On Tue, 30 Oct 2007 09:56:08 -0700,
> > Dirk Hohndel <hohndel@linux.intel.com> wrote:
> > 
> > > > > IIRC, Al recently vetoed a similar patch. As far as I'm concerned, with
> > > > > the correct return values, the patch then looks fine to me.
> 
> So Al, are you ok with this one?
> 
> > > > We need some kind of check concerning the kobject to avoid mysterious
> > > > errors (especially checking for the failed kobject_add() is needed).
> > > > Whether we want just to inform the user of the failure instead of
> > > > failing the function is another question.
> > > 
> > > What are you suggesting? I'd love to make the behaviour consistent everywhere
> > > (and am willing to go through things in order to make that happen), but what is
> > > the consistent behaviour that we'd want?
> > 
> > I'd be fine with just propagating the error after cleanup (that is what
> > for example the driver core usually does), but I don't know the
> > surrounding code well enough for a definitive answer.
> 
> Ok, I think I have it consistent now. I also ran it through checkpatch.pl :-)
> 
> /D
> 
> 
> [FILESYSTEM] add_partition ignores errors

Looks good to me. One final return value note:

> @@ -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;
>  		}

-EBUSY seems a bit confusing here, although I don't know what the best
value to return would be (and it probably doesn't matter). -EOVERFLOW?
-ENOSPC?

-- 
Jens Axboe