[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <YmhhcQuXO7ZKjolH@casper.infradead.org>
Date: Tue, 26 Apr 2022 22:17:37 +0100
From: Matthew Wilcox <willy@...radead.org>
To: Andrew Morton <akpm@...ux-foundation.org>
Cc: Wan Jiabing <wanjiabing@...o.com>,
Vlastimil Babka <vbabka@...e.cz>,
David Howells <dhowells@...hat.com>, linux-mm@...ck.org,
linux-kernel@...r.kernel.org, Julia Lawall <Julia.Lawall@...ia.fr>
Subject: Re: [PATCH] mm/filemap: Fix NULL pointer dereference in
pagecache_get_page
On Tue, Apr 26, 2022 at 02:06:41PM -0700, Andrew Morton wrote:
> On Tue, 26 Apr 2022 13:08:08 +0100 Matthew Wilcox <willy@...radead.org> wrote:
>
> > On Tue, Apr 26, 2022 at 04:25:48PM +0800, Wan Jiabing wrote:
> > > Fix following coccicheck error:
> > > mm/folio-compat.c:128:17-21: ERROR: folio is NULL but dereferenced.
> > > folio = __filemap_get_folio(mapping, index, fgp_flags, gfp);
> > > - if ((fgp_flags & FGP_HEAD) || !folio || xa_is_value(folio))
> > > + if (!folio)
> > > + return NULL;
> > > + if ((fgp_flags & FGP_HEAD) || xa_is_value(folio))
> > > return &folio->page;
> >
> > That doesn't dereference the folio. Coccicheck is wrong.
>
> Doing
>
> return &(0->page);
>
> is a rather obscure way of doing `return NULL;'.
>
> I agree the patch doesn't fix anything, but it results in saner-looking code?
I suppose that's in the eye of the beholder? The original code makes
more sense to me. Besides, it's in the folio-compat file; nobody
should be looking at that except to figure out "What function should I
be transitioning to?"
Powered by blists - more mailing lists