From: Greg Kurz <groug@kaod.org>
To: Stefan Hajnoczi <stefanha@redhat.com>
Cc: mszeredi@redhat.com, Daniel Berrange <berrange@redhat.com>,
slp@redhat.com, qemu-devel@nongnu.org,
"Dr. David Alan Gilbert" <dgilbert@redhat.com>,
virtio-fs@redhat.com, P J P <ppandit@redhat.com>,
Laszlo Ersek <lersek@redhat.com>,
vgoyal@redhat.com
Subject: Re: [PATCH v4 2/3] virtiofsd: optionally return inode pointer from lo_do_lookup()
Date: Thu, 4 Feb 2021 12:19:27 +0100 [thread overview]
Message-ID: <20210204121927.31be17ff@bahia.lan> (raw)
In-Reply-To: <20210204094537.GB344659@stefanha-x1.localdomain>
[-- Attachment #1: Type: text/plain, Size: 4543 bytes --]
On Thu, 4 Feb 2021 09:45:37 +0000
Stefan Hajnoczi <stefanha@redhat.com> wrote:
> On Thu, Feb 04, 2021 at 09:25:28AM +0100, Greg Kurz wrote:
> > On Wed, 3 Feb 2021 17:00:06 +0000
> > Stefan Hajnoczi <stefanha@redhat.com> wrote:
> >
> > > On Wed, Feb 03, 2021 at 03:20:14PM +0100, Greg Kurz wrote:
> > > > On Wed, 3 Feb 2021 11:37:18 +0000
> > > > Stefan Hajnoczi <stefanha@redhat.com> wrote:
> > > >
> > > > > lo_do_lookup() finds an existing inode or allocates a new one. It
> > > > > increments nlookup so that the inode stays alive until the client
> > > > > releases it.
> > > > >
> > > > > Existing callers don't need the struct lo_inode so the function doesn't
> > > > > return it. Extend the function to optionally return the inode. The next
> > > > > commit will need it.
> > > > >
> > > > > Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
> > > > > ---
> > > > > tools/virtiofsd/passthrough_ll.c | 29 +++++++++++++++++++++--------
> > > > > 1 file changed, 21 insertions(+), 8 deletions(-)
> > > > >
> > > > > diff --git a/tools/virtiofsd/passthrough_ll.c b/tools/virtiofsd/passthrough_ll.c
> > > > > index e63cbd3fb7..c87a1f3d72 100644
> > > > > --- a/tools/virtiofsd/passthrough_ll.c
> > > > > +++ b/tools/virtiofsd/passthrough_ll.c
> > > > > @@ -831,11 +831,13 @@ static int do_statx(struct lo_data *lo, int dirfd, const char *pathname,
> > > > > }
> > > > >
> > > > > /*
> > > > > - * Increments nlookup and caller must release refcount using
> > > > > - * lo_inode_put(&parent).
> > > > > + * Increments nlookup on the inode on success. unref_inode_lolocked() must be
> > > > > + * called eventually to decrement nlookup again. If inodep is non-NULL, the
> > > > > + * inode pointer is stored and the caller must call lo_inode_put().
> > > > > */
> > > > > static int lo_do_lookup(fuse_req_t req, fuse_ino_t parent, const char *name,
> > > > > - struct fuse_entry_param *e)
> > > > > + struct fuse_entry_param *e,
> > > > > + struct lo_inode **inodep)
> > > > > {
> > > > > int newfd;
> > > > > int res;
> > > > > @@ -845,6 +847,10 @@ static int lo_do_lookup(fuse_req_t req, fuse_ino_t parent, const char *name,
> > > > > struct lo_inode *inode = NULL;
> > > > > struct lo_inode *dir = lo_inode(req, parent);
> > > > >
> > > > > + if (inodep) {
> > > > > + *inodep = NULL;
> > > > > + }
> > > > > +
> > > >
> > > > Is this side-effect needed ? If lo_do_lookup() returns an error, it
> > > > rather seems that the caller shouldn't expect anything to be written
> > > > here, i.e. the content of *inodep still belongs to the caller and
> > > > whatever value it previously put in there (as patch 3/3 does) should
> > > > be preserved IMHO.
> > > >
> > > > Apart from that LGTM.
> > >
> > > I like this approach because it prevents accessing uninitialized memory
> > > in the caller:
> > >
> > > struct lo_inode *inode;
> > >
> > > if (lo_do_lookup(..., &inodep) != 0) {
> > > goto err;
> > > }
> > > ...
> > >
> > > err:
> > > lo_inode_put(&inode); <-- uninitialized in the error case!
> >
> > My point is that it is the caller's business to ensure that inode
> > doesn't contain garbage if it is to be used irrespective of the
> > outcome of lo_do_lookup(). This is precisely what patch 3/3 does,
> > so I don't understand the ultimate purpose of nullifying the
> > inode pointer _again_ in lo_do_lookup()...
>
> APIs should be designed to eliminate classes of errors where possible
> IMO. Taking care regarding the uninitialized pointer in the error case
> could be the caller's responsibility, but what's the advantage?
>
Because it is more explicit. FWIW caller is still responsible since it
calls lo_inode_put() in the end : initializing inode to NULL like patch
3/3 does warrants that no matter what happens in lo_do_lookup() or even
if it isn't called at all, inode can be safely passed to lo_inode_put().
But this change doesn't hurt, especially with the benefits of the rest
of this series, so:
Reviewed-by: Greg Kurz <groug@kaod.org>
> (There's a related thing with lo_inode_put(&inode) where it sets *inode
> = NULL to eliminate use-after-free bugs in callers. It would have been
> possible to use the same approach as free(3) where it's the caller's
> responsiblity, but that API design decision in free(3) has caused
> many bugs in applications.)
>
> Stefan
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
next prev parent reply other threads:[~2021-02-04 11:21 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-02-03 11:37 [PATCH v4 0/3] virtiofsd: prevent opening of special files (CVE-2020-35517) Stefan Hajnoczi
2021-02-03 11:37 ` [PATCH v4 1/3] virtiofsd: extract lo_do_open() from lo_open() Stefan Hajnoczi
2021-02-03 14:20 ` Greg Kurz
2021-02-03 14:47 ` Dr. David Alan Gilbert
2021-02-03 15:45 ` Greg Kurz
2021-02-03 17:47 ` Greg Kurz
2021-02-03 16:57 ` Stefan Hajnoczi
2021-02-03 11:37 ` [PATCH v4 2/3] virtiofsd: optionally return inode pointer from lo_do_lookup() Stefan Hajnoczi
2021-02-03 14:20 ` Greg Kurz
2021-02-03 17:00 ` Stefan Hajnoczi
2021-02-04 8:25 ` Greg Kurz
2021-02-04 9:45 ` Stefan Hajnoczi
2021-02-04 11:19 ` Greg Kurz [this message]
2021-02-03 11:37 ` [PATCH v4 3/3] virtiofsd: prevent opening of special files (CVE-2020-35517) Stefan Hajnoczi
2021-02-03 15:28 ` Vivek Goyal
2021-02-03 16:02 ` Greg Kurz
2021-02-03 16:08 ` Vivek Goyal
2021-02-03 17:05 ` Stefan Hajnoczi
2021-02-03 18:05 ` Dr. David Alan Gilbert
2021-02-03 21:14 ` Vivek Goyal
2021-02-04 9:47 ` Stefan Hajnoczi
2021-02-03 15:57 ` Greg Kurz
2021-02-03 17:06 ` Stefan Hajnoczi
2021-02-03 11:46 ` [PATCH v4 0/3] " no-reply
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20210204121927.31be17ff@bahia.lan \
--to=groug@kaod.org \
--cc=berrange@redhat.com \
--cc=dgilbert@redhat.com \
--cc=lersek@redhat.com \
--cc=mszeredi@redhat.com \
--cc=ppandit@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=slp@redhat.com \
--cc=stefanha@redhat.com \
--cc=vgoyal@redhat.com \
--cc=virtio-fs@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).