Re: [PATCH 6/6] ksm: unstable_tree_search_insert error checking cleanup

[Hugh Dickins <hughd@google.com>]

On Mon, 2 Nov 2015, Andrea Arcangeli wrote:
> On Sun, Oct 25, 2015 at 04:18:05PM -0700, Hugh Dickins wrote:
> > On Thu, 15 Oct 2015, Andrea Arcangeli wrote:
> > 
> > > get_mergeable_page() can only return NULL (in case of errors) or the
> > > pinned mergeable page. It can't return an error different than
> > > NULL. This makes it more readable and less confusion in addition to
> > > optimizing the check.
> > > 
> > > Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
> > 
> > I share your sentiment, prefer to avoid an unnecessary IS_ERR_OR_NULL.
> > And you may be right that it's unnecessary; but that's far from clear
> > to me, and you haven't changed the IS_ERR_OR_NULL after follow_page()
> > in get_mergeable_page() where it originates, so I wonder if you just
> > got confused on this.
> > 
> > Even if you have established that there's currently no way that
> > follow_page(vma, addr, FOLL_GET) could return an -errno on a vma
> > validated by find_mergeable_vma(), I think we'd still be better off
> > to allow for some future -errno there; but I'd be happy for you to
> > keep the change below, but also adjust get_mergeable_page() to
> > convert an -errno immediately to NULL after follow_page().
> > 
> > So, I think this is gently Nacked in its present form,
> > but a replacement eagerly Acked.
> 
> The "out:" label is followed by page = NULL, so if follow_page returns
> an error, get_mergeable_page still cannot return an error.

Ah, yes, I missed that, thanks.

Acked-by: Hugh Dickins <hughd@google.com>

> 
> If this wasn't the case, get_mergeable_page would return an
> uninitialized pointer if find_mergeable_vma would return NULL.
> 
> I guess the IS_ERR_OR_NULL that I removed, was the direct result of
> overlooking the location of the "out:".
> 
> If there was a return after the "out:" the readability would have been
> improved, but I haven't touched the code around the "out:". That's the
> way it was before. Now I'll add a return in this same patch after the
> "out:" before resubmitting without 1/6.

I can't very well claim now that it was readable before, can I?-)
We'll see; I wouldn't mind if you leave this patch as is.

Hugh

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>