All of lore.kernel.org
 help / color / mirror / Atom feed
From: Stefan Hajnoczi <stefanha@redhat.com>
To: Vivek Goyal <vgoyal@redhat.com>
Cc: mszeredi@redhat.com, slp@redhat.com, qemu-devel@nongnu.org,
	P J P <ppandit@redhat.com>,
	virtio-fs@redhat.com, Alex Xu <alex@alxu.ca>,
	"Dr. David Alan Gilbert" <dgilbert@redhat.com>
Subject: Re: [PATCH] virtiofsd: prevent opening of special files (CVE-2020-35517)
Date: Tue, 26 Jan 2021 10:10:31 +0000	[thread overview]
Message-ID: <20210126101031.GB226981@stefanha-x1.localdomain> (raw)
In-Reply-To: <20210122154054.GA120574@redhat.com>

[-- Attachment #1: Type: text/plain, Size: 2235 bytes --]

On Fri, Jan 22, 2021 at 10:40:54AM -0500, Vivek Goyal wrote:
> On Thu, Jan 21, 2021 at 02:44:29PM +0000, Stefan Hajnoczi wrote:
> > A well-behaved FUSE client does not attempt to open special files with
> > FUSE_OPEN because they are handled on the client side (e.g. device nodes
> > are handled by client-side device drivers).
> > 
> > The check to prevent virtiofsd from opening special files is missing in
> > a few cases, most notably FUSE_OPEN. A malicious client can cause
> > virtiofsd to open a device node, potentially allowing the guest to
> > escape. This can be exploited by a modified guest device driver. It is
> > not exploitable from guest userspace since the guest kernel will handle
> > special files inside the guest instead of sending FUSE requests.
> > 
> > This patch adds the missing checks to virtiofsd. This is a short-term
> > solution because it does not prevent a compromised virtiofsd process
> > from opening device nodes on the host.
> > 
> > Reported-by: Alex Xu <alex@alxu.ca>
> > Fixes: CVE-2020-35517
> > Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
> 
> It looks good to me. I see there is another openat() instance in
> lo_opendir(). May be we can convert it to use lo_inode_open() as well?

I thought so too, but this one is interesting:

  fd = openat(lo_fd(req, ino), ".", O_RDONLY);

Using "." here is basically equivalent to O_DIRECTORY!

Therefore this always fails on special files. That's why I ended up
leaving it unchanged.

  $ cat a.c
  #define _GNU_SOURCE
  #include <sys/types.h>
  #include <sys/stat.h>
  #include <fcntl.h>
  #include <unistd.h>
  #include <stdio.h>
  #include <stdlib.h>

  int main(int argc, char **argv)
  {
      int fd;
      int pfd = open(argv[1], O_PATH);

      if (pfd < 0) {
          perror("open");
          return EXIT_FAILURE;
      }

      fd = openat(pfd, ".", O_RDONLY);
      if (fd < 0) {
          perror("openat");
          return EXIT_FAILURE;
      }

      close(fd);
      close(pfd);
      return EXIT_SUCCESS;
  }
  $ gcc -o a a.c
  $ ./a /tmp
  <--- no error
  $ ./a /tmp/a.c
  openat: Not a directory
  $ ./a /dev/stdin
  openat: Not a directory

Stefan

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

WARNING: multiple messages have this Message-ID (diff)
From: Stefan Hajnoczi <stefanha@redhat.com>
To: Vivek Goyal <vgoyal@redhat.com>
Cc: qemu-devel@nongnu.org, P J P <ppandit@redhat.com>,
	virtio-fs@redhat.com, Alex Xu <alex@alxu.ca>
Subject: Re: [Virtio-fs] [PATCH] virtiofsd: prevent opening of special files (CVE-2020-35517)
Date: Tue, 26 Jan 2021 10:10:31 +0000	[thread overview]
Message-ID: <20210126101031.GB226981@stefanha-x1.localdomain> (raw)
In-Reply-To: <20210122154054.GA120574@redhat.com>

[-- Attachment #1: Type: text/plain, Size: 2235 bytes --]

On Fri, Jan 22, 2021 at 10:40:54AM -0500, Vivek Goyal wrote:
> On Thu, Jan 21, 2021 at 02:44:29PM +0000, Stefan Hajnoczi wrote:
> > A well-behaved FUSE client does not attempt to open special files with
> > FUSE_OPEN because they are handled on the client side (e.g. device nodes
> > are handled by client-side device drivers).
> > 
> > The check to prevent virtiofsd from opening special files is missing in
> > a few cases, most notably FUSE_OPEN. A malicious client can cause
> > virtiofsd to open a device node, potentially allowing the guest to
> > escape. This can be exploited by a modified guest device driver. It is
> > not exploitable from guest userspace since the guest kernel will handle
> > special files inside the guest instead of sending FUSE requests.
> > 
> > This patch adds the missing checks to virtiofsd. This is a short-term
> > solution because it does not prevent a compromised virtiofsd process
> > from opening device nodes on the host.
> > 
> > Reported-by: Alex Xu <alex@alxu.ca>
> > Fixes: CVE-2020-35517
> > Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
> 
> It looks good to me. I see there is another openat() instance in
> lo_opendir(). May be we can convert it to use lo_inode_open() as well?

I thought so too, but this one is interesting:

  fd = openat(lo_fd(req, ino), ".", O_RDONLY);

Using "." here is basically equivalent to O_DIRECTORY!

Therefore this always fails on special files. That's why I ended up
leaving it unchanged.

  $ cat a.c
  #define _GNU_SOURCE
  #include <sys/types.h>
  #include <sys/stat.h>
  #include <fcntl.h>
  #include <unistd.h>
  #include <stdio.h>
  #include <stdlib.h>

  int main(int argc, char **argv)
  {
      int fd;
      int pfd = open(argv[1], O_PATH);

      if (pfd < 0) {
          perror("open");
          return EXIT_FAILURE;
      }

      fd = openat(pfd, ".", O_RDONLY);
      if (fd < 0) {
          perror("openat");
          return EXIT_FAILURE;
      }

      close(fd);
      close(pfd);
      return EXIT_SUCCESS;
  }
  $ gcc -o a a.c
  $ ./a /tmp
  <--- no error
  $ ./a /tmp/a.c
  openat: Not a directory
  $ ./a /dev/stdin
  openat: Not a directory

Stefan

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

  reply	other threads:[~2021-01-26 10:12 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-01-21 14:44 [PATCH] virtiofsd: prevent opening of special files (CVE-2020-35517) Stefan Hajnoczi
2021-01-21 14:44 ` [Virtio-fs] " Stefan Hajnoczi
2021-01-21 14:48 ` Daniel P. Berrangé
2021-01-21 14:48   ` [Virtio-fs] " Daniel P. Berrangé
2021-01-22 15:49   ` Vivek Goyal
2021-01-22 15:49     ` [Virtio-fs] " Vivek Goyal
2021-01-25 14:53   ` Stefan Hajnoczi
2021-01-25 14:53     ` [Virtio-fs] " Stefan Hajnoczi
2021-01-21 14:50 ` alex--- via
2021-01-21 14:50   ` [Virtio-fs] " Alex Xu
2021-01-21 15:32 ` Laszlo Ersek
2021-01-21 15:32   ` [Virtio-fs] " Laszlo Ersek
2021-01-21 15:52   ` alex--- via
2021-01-21 15:52     ` [Virtio-fs] " Alex Xu
2021-01-21 17:07     ` Laszlo Ersek
2021-01-21 17:07       ` [Virtio-fs] " Laszlo Ersek
2021-01-21 19:00 ` Dr. David Alan Gilbert
2021-01-21 19:00   ` [Virtio-fs] " Dr. David Alan Gilbert
2021-01-22 15:40 ` Vivek Goyal
2021-01-22 15:40   ` [Virtio-fs] " Vivek Goyal
2021-01-26 10:10   ` Stefan Hajnoczi [this message]
2021-01-26 10:10     ` Stefan Hajnoczi
2021-01-25 16:12 ` Miklos Szeredi
2021-01-25 16:12   ` [Virtio-fs] " Miklos Szeredi
2021-01-26 10:18   ` Stefan Hajnoczi
2021-01-26 10:18     ` [Virtio-fs] " Stefan Hajnoczi
2021-01-26 10:27     ` Miklos Szeredi
2021-01-26 10:27       ` [Virtio-fs] " Miklos Szeredi

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=20210126101031.GB226981@stefanha-x1.localdomain \
    --to=stefanha@redhat.com \
    --cc=alex@alxu.ca \
    --cc=dgilbert@redhat.com \
    --cc=mszeredi@redhat.com \
    --cc=ppandit@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=slp@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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.