Re: [PATCH v3 0/3] last batch of add_disk() error handling conversions

[Luis Chamberlain <mcgrof@kernel.org>]

On Fri, Oct 22, 2021 at 10:08:39AM -0700, Luis Chamberlain wrote:
> On Fri, Oct 22, 2021 at 10:06:07AM +0900, Tetsuo Handa wrote:
> > On 2021/10/22 1:38, Luis Chamberlain wrote:
> > > I rebased Tetsuo Handa's patch onto the latest linux-next as this
> > > series depends on it, and so I am sending it part of this series as
> > > without it, this won't apply. Tetsuo, does the rebase of your patch
> > > look OK?
> > 
> > OK, though I wanted my fix to be sent to upstream and stable before this series.
> 
> Sure, absolutely, your patch is certainly separate and is needed as a
> fix downstream to stable it would seem.
> 
> > > If it is not too much trouble, I'd like to ask for testing for the
> > > ataflop changes from Michael Schmitz, if possible, that is he'd just
> > > have to merge Tetsuo's rebased patch and the 2nd patch in this series.
> > > This is all rebased on linux-next tag 20211020.
> > 
> > Yes, please.
> > 
> > After this series, I guess we can remove "bool registered[NUM_DISK_MINORS];" like below
> > due to (unit[drive].disk[type] != NULL) == (unit[drive].registered[type] == true).
> 
> Sounds good.
> 
> > Regarding this series, setting unit[drive].registered[type] = true in ataflop_probe() is
> > pointless because atari_floppy_cleanup() checks unit[i].disk[type] != NULL for calling
> > del_gendisk().
> 
> I see, will fix.

Actually just not doing it for that case seems odd, so I would prefer
the removal of the bool registered to be a separate patch to make it
clearer.

  Luis