Re: [syzbot] KMSAN: uninit-value in pagecache_write

From: Alexander Potapenko <glider@google.com>
To: Eric Biggers <ebiggers@kernel.org>
Cc: syzbot <syzbot+9767be679ef5016b6082@syzkaller.appspotmail.com>,
	adilger.kernel@dilger.ca, linux-ext4@vger.kernel.org,
	linux-kernel@vger.kernel.org, syzkaller-bugs@googlegroups.com,
	tytso@mit.edu
Subject: Re: [syzbot] KMSAN: uninit-value in pagecache_write
Date: Thu, 10 Nov 2022 12:01:13 +0100	[thread overview]
Message-ID: <CAG_fn=WVUmodhyY+bNyKWC6BM5gXaq3fguw70ngsjZWc4mXkPg@mail.gmail.com> (raw)
In-Reply-To: <Y2qUylpsZcJ7HF0Z@sol.localdomain>

On Tue, Nov 8, 2022 at 6:41 PM Eric Biggers <ebiggers@kernel.org> wrote:
>
> On Tue, Nov 08, 2022 at 10:08:36AM +0100, 'Alexander Potapenko' via syzkaller-bugs wrote:
> > > >
> > > > Anyway, this patch doesn't hurt, I suppose.  Can please you send it out as a
> > > > formal patch to linux-ext4?  It would be easy for people to miss this patch
> > > > buried in this thread.  Also, can you please send a patch to linux-f2fs-devel
> > > > for the same code in fs/f2fs/verity.c?
> > >
> > > Will do!
> >
> > Shall I also initialize fsdata here:
> >
> > $ git grep 'void \*fsdata;'
> > fs/affs/file.c:         void *fsdata;
> > fs/ext4/verity.c:               void *fsdata;
> > fs/f2fs/verity.c:               void *fsdata;
> > fs/hfs/extent.c:                void *fsdata;
> > fs/hfsplus/extents.c:           void *fsdata;
> > fs/ocfs2/mmap.c:        void *fsdata;
>
> Yes, it looks like they all need this.  Except maybe ocfs2?  It's hard to tell.

For ocfs2 the begin/end functions are always the same, so it's harder
to mess fsdata up.
Guess we can say for now that __ocfs2_page_mkwrite() never passes an
uninitialized variable to another function.

> - Eric

-- 
Alexander Potapenko
Software Engineer

Google Germany GmbH
Erika-Mann-Straße, 33
80636 München

Geschäftsführer: Paul Manicle, Liana Sebastian
Registergericht und -nummer: Hamburg, HRB 86891
Sitz der Gesellschaft: Hamburg