[v3,1/3] VirtIO-RNG: Update default entropy source to `/dev/urandom`

Commit Message

Laurent Vivier May 10, 2019, 1:42 p.m. UTC

From: Kashyap Chamarthy <[e-mail]>

When QEMU exposes a VirtIO-RNG device to the guest, that device needs a
source of entropy, and that source needs to be "non-blocking", like
`/dev/urandom`.  However, currently QEMU defaults to the problematic
`/dev/random`, which is "blocking" (as in, it waits until sufficient
entropy is available).

Why prefer `/dev/urandom` over `/dev/random`?
---------------------------------------------

The man pages of urandom(4) and random(4) state:

    "The /dev/random device is a legacy interface which dates back to a
    time where the cryptographic primitives used in the implementation
    of /dev/urandom were not widely trusted.  It will return random
    bytes only within the estimated number of bits of fresh noise in the
    entropy pool, blocking if necessary.  /dev/random is suitable for
    applications that need high quality randomness, and can afford
    indeterminate delays."

Further, the "Usage" section of the said man pages state:

    "The /dev/random interface is considered a legacy interface, and
    /dev/urandom is preferred and sufficient in all use cases, with the
    exception of applications which require randomness during early boot
    time; for these applications, getrandom(2) must be used instead,
    because it will block until the entropy pool is initialized.

    "If a seed file is saved across reboots as recommended below (all
    major Linux distributions have done this since 2000 at least), the
    output is cryptographically secure against attackers without local
    root access as soon as it is reloaded in the boot sequence, and
    perfectly adequate for network encryption session keys.  Since reads
    from /dev/random may block, users will usually want to open it in
    nonblocking mode (or perform a read with timeout), and provide some
    sort of user notification if the desired entropy is not immediately
    available."

And refer to random(7) for a comparison of `/dev/random` and
`/dev/urandom`.

    - - -

Given the above, change the entropy source for VirtIO-RNG device to
`/dev/urandom`.

Related discussion in these[1][2] past threads.

[1] https://lists.nongnu.org/archive/html/qemu-devel/2018-06/msg08335.html
    -- "RNG: Any reason QEMU doesn't default to `/dev/urandom`?"
[2] https://lists.nongnu.org/archive/html/qemu-devel/2018-09/msg02724.html
    -- "[RFC] Virtio RNG: Consider changing the default entropy source to
       /dev/urandom"

Signed-off-by: Kashyap Chamarthy <[e-mail]>
Reviewed-by: Daniel P. Berrangé <[e-mail]>
Reviewed-by: Stefan Hajnoczi <[e-mail]>
Signed-off-by: Laurent Vivier <[e-mail]>
---
 backends/rng-random.c | 2 +-
 qemu-options.hx       | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

Comments

Michael S. Tsirkin May 10, 2019, 4:12 p.m. UTC | #1

On Fri, May 10, 2019 at 03:42:01PM +0200, Laurent Vivier wrote:
> From: Kashyap Chamarthy <[e-mail]>
> 
> When QEMU exposes a VirtIO-RNG device to the guest, that device needs a
> source of entropy, and that source needs to be "non-blocking", like
> `/dev/urandom`.  However, currently QEMU defaults to the problematic
> `/dev/random`, which is "blocking" (as in, it waits until sufficient
> entropy is available).
> 
> Why prefer `/dev/urandom` over `/dev/random`?
> ---------------------------------------------
> 
> The man pages of urandom(4) and random(4) state:
> 
>     "The /dev/random device is a legacy interface which dates back to a
>     time where the cryptographic primitives used in the implementation
>     of /dev/urandom were not widely trusted.  It will return random
>     bytes only within the estimated number of bits of fresh noise in the
>     entropy pool, blocking if necessary.  /dev/random is suitable for
>     applications that need high quality randomness, and can afford
>     indeterminate delays."
> 
> Further, the "Usage" section of the said man pages state:
> 
>     "The /dev/random interface is considered a legacy interface, and
>     /dev/urandom is preferred and sufficient in all use cases, with the
>     exception of applications which require randomness during early boot
>     time; for these applications, getrandom(2) must be used instead,
>     because it will block until the entropy pool is initialized.

So how about just using getrandom then?

> 
>     "If a seed file is saved across reboots as recommended below (all
>     major Linux distributions have done this since 2000 at least), the
>     output is cryptographically secure against attackers without local
>     root access as soon as it is reloaded in the boot sequence, and
>     perfectly adequate for network encryption session keys.  Since reads
>     from /dev/random may block, users will usually want to open it in
>     nonblocking mode (or perform a read with timeout), and provide some
>     sort of user notification if the desired entropy is not immediately
>     available."
> 
> And refer to random(7) for a comparison of `/dev/random` and
> `/dev/urandom`.
> 
>     - - -
> 
> Given the above, change the entropy source for VirtIO-RNG device to
> `/dev/urandom`.
> 
> Related discussion in these[1][2] past threads.
> 
> [1] https://lists.nongnu.org/archive/html/qemu-devel/2018-06/msg08335.html
>     -- "RNG: Any reason QEMU doesn't default to `/dev/urandom`?"
> [2] https://lists.nongnu.org/archive/html/qemu-devel/2018-09/msg02724.html
>     -- "[RFC] Virtio RNG: Consider changing the default entropy source to
>        /dev/urandom"
> 
> Signed-off-by: Kashyap Chamarthy <[e-mail]>
> Reviewed-by: Daniel P. Berrangé <[e-mail]>
> Reviewed-by: Stefan Hajnoczi <[e-mail]>
> Signed-off-by: Laurent Vivier <[e-mail]>
> ---
>  backends/rng-random.c | 2 +-
>  qemu-options.hx       | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/backends/rng-random.c b/backends/rng-random.c
> index e2a49b0571d7..eff36ef14084 100644
> --- a/backends/rng-random.c
> +++ b/backends/rng-random.c
> @@ -112,7 +112,7 @@ static void rng_random_init(Object *obj)
>                              rng_random_set_filename,
>                              NULL);
>  
> -    s->filename = g_strdup("/dev/random");
> +    s->filename = g_strdup("/dev/urandom");
>      s->fd = -1;
>  }
>  
> diff --git a/qemu-options.hx b/qemu-options.hx
> index 0191ef8b1eb7..4df0ea3aed5c 100644
> --- a/qemu-options.hx
> +++ b/qemu-options.hx
> @@ -4286,7 +4286,7 @@ Creates a random number generator backend which obtains entropy from
>  a device on the host. The @option{id} parameter is a unique ID that
>  will be used to reference this entropy backend from the @option{virtio-rng}
>  device. The @option{filename} parameter specifies which file to obtain
> -entropy from and if omitted defaults to @option{/dev/random}.
> +entropy from and if omitted defaults to @option{/dev/urandom}.
>  
>  @item -object rng-egd,id=@var{id},chardev=@var{chardevid}
>  
> -- 
> 2.20.1

Daniel P. Berrangé May 10, 2019, 4:16 p.m. UTC | #2

On Fri, May 10, 2019 at 12:12:41PM -0400, Michael S. Tsirkin wrote:
> On Fri, May 10, 2019 at 03:42:01PM +0200, Laurent Vivier wrote:
> > From: Kashyap Chamarthy <[e-mail]>
> > 
> > When QEMU exposes a VirtIO-RNG device to the guest, that device needs a
> > source of entropy, and that source needs to be "non-blocking", like
> > `/dev/urandom`.  However, currently QEMU defaults to the problematic
> > `/dev/random`, which is "blocking" (as in, it waits until sufficient
> > entropy is available).
> > 
> > Why prefer `/dev/urandom` over `/dev/random`?
> > ---------------------------------------------
> > 
> > The man pages of urandom(4) and random(4) state:
> > 
> >     "The /dev/random device is a legacy interface which dates back to a
> >     time where the cryptographic primitives used in the implementation
> >     of /dev/urandom were not widely trusted.  It will return random
> >     bytes only within the estimated number of bits of fresh noise in the
> >     entropy pool, blocking if necessary.  /dev/random is suitable for
> >     applications that need high quality randomness, and can afford
> >     indeterminate delays."
> > 
> > Further, the "Usage" section of the said man pages state:
> > 
> >     "The /dev/random interface is considered a legacy interface, and
> >     /dev/urandom is preferred and sufficient in all use cases, with the
> >     exception of applications which require randomness during early boot
> >     time; for these applications, getrandom(2) must be used instead,
> >     because it will block until the entropy pool is initialized.
> 
> So how about just using getrandom then?

The 3rd patch in this series addresses that.


Regards,
Daniel

Markus Armbruster May 10, 2019, 4:18 p.m. UTC | #3

Laurent Vivier <[e-mail]> writes:

> From: Kashyap Chamarthy <[e-mail]>
>
> When QEMU exposes a VirtIO-RNG device to the guest, that device needs a
> source of entropy, and that source needs to be "non-blocking", like
> `/dev/urandom`.  However, currently QEMU defaults to the problematic
> `/dev/random`, which is "blocking" (as in, it waits until sufficient
> entropy is available).
>
> Why prefer `/dev/urandom` over `/dev/random`?
> ---------------------------------------------
>
> The man pages of urandom(4) and random(4) state:
>
>     "The /dev/random device is a legacy interface which dates back to a
>     time where the cryptographic primitives used in the implementation
>     of /dev/urandom were not widely trusted.  It will return random
>     bytes only within the estimated number of bits of fresh noise in the
>     entropy pool, blocking if necessary.  /dev/random is suitable for
>     applications that need high quality randomness, and can afford
>     indeterminate delays."
>
> Further, the "Usage" section of the said man pages state:
>
>     "The /dev/random interface is considered a legacy interface, and
>     /dev/urandom is preferred and sufficient in all use cases, with the
>     exception of applications which require randomness during early boot
>     time; for these applications, getrandom(2) must be used instead,
>     because it will block until the entropy pool is initialized.
>
>     "If a seed file is saved across reboots as recommended below (all
>     major Linux distributions have done this since 2000 at least), the
>     output is cryptographically secure against attackers without local
>     root access as soon as it is reloaded in the boot sequence, and
>     perfectly adequate for network encryption session keys.  Since reads
>     from /dev/random may block, users will usually want to open it in
>     nonblocking mode (or perform a read with timeout), and provide some
>     sort of user notification if the desired entropy is not immediately
>     available."
>
> And refer to random(7) for a comparison of `/dev/random` and
> `/dev/urandom`.
>
>     - - -
>
> Given the above, change the entropy source for VirtIO-RNG device to
> `/dev/urandom`.
>
> Related discussion in these[1][2] past threads.
>
> [1] https://lists.nongnu.org/archive/html/qemu-devel/2018-06/msg08335.html
>     -- "RNG: Any reason QEMU doesn't default to `/dev/urandom`?"
> [2] https://lists.nongnu.org/archive/html/qemu-devel/2018-09/msg02724.html
>     -- "[RFC] Virtio RNG: Consider changing the default entropy source to
>        /dev/urandom"
>
> Signed-off-by: Kashyap Chamarthy <[e-mail]>
> Reviewed-by: Daniel P. Berrangé <[e-mail]>
> Reviewed-by: Stefan Hajnoczi <[e-mail]>
> Signed-off-by: Laurent Vivier <[e-mail]>

Suggest to work Daniel's analysis into the commit message, like this:

  When QEMU exposes a VirtIO-RNG device to the guest, that device needs a
  source of entropy, and that source needs to be "non-blocking", like
  `/dev/urandom`.  However, currently QEMU defaults to the problematic
  `/dev/random`, which on Linux is "blocking" (as in, it waits until
  sufficient entropy is available).

  Why prefer `/dev/urandom` over `/dev/random` on Linux?
  ------------------------------------------------------
  [...]

  What about other OSes?
  ----------------------

  /dev/urandom exists and works on OS-X, FreeBSD, DragonFlyBSD, NetBSD
  and OpenBSD, which covers all the non-Linux platforms we explicitly
  support, aside from Windows.

  On Windows /dev/random doesn't work either so we don't regress. This
  is actually another argument in favour of using the newly proposed
  rng-builtin backend by default, as that will work on Windows.

      - - -

  Given the above, change the entropy source for VirtIO-RNG device to
  `/dev/urandom`.
  [...]

Michael S. Tsirkin May 10, 2019, 4:21 p.m. UTC | #4

On Fri, May 10, 2019 at 05:16:44PM +0100, Daniel P. Berrangé wrote:
> On Fri, May 10, 2019 at 12:12:41PM -0400, Michael S. Tsirkin wrote:
> > On Fri, May 10, 2019 at 03:42:01PM +0200, Laurent Vivier wrote:
> > > From: Kashyap Chamarthy <[e-mail]>
> > > 
> > > When QEMU exposes a VirtIO-RNG device to the guest, that device needs a
> > > source of entropy, and that source needs to be "non-blocking", like
> > > `/dev/urandom`.  However, currently QEMU defaults to the problematic
> > > `/dev/random`, which is "blocking" (as in, it waits until sufficient
> > > entropy is available).
> > > 
> > > Why prefer `/dev/urandom` over `/dev/random`?
> > > ---------------------------------------------
> > > 
> > > The man pages of urandom(4) and random(4) state:
> > > 
> > >     "The /dev/random device is a legacy interface which dates back to a
> > >     time where the cryptographic primitives used in the implementation
> > >     of /dev/urandom were not widely trusted.  It will return random
> > >     bytes only within the estimated number of bits of fresh noise in the
> > >     entropy pool, blocking if necessary.  /dev/random is suitable for
> > >     applications that need high quality randomness, and can afford
> > >     indeterminate delays."
> > > 
> > > Further, the "Usage" section of the said man pages state:
> > > 
> > >     "The /dev/random interface is considered a legacy interface, and
> > >     /dev/urandom is preferred and sufficient in all use cases, with the
> > >     exception of applications which require randomness during early boot
> > >     time; for these applications, getrandom(2) must be used instead,
> > >     because it will block until the entropy pool is initialized.
> > 
> > So how about just using getrandom then?
> 
> The 3rd patch in this series addresses that.

It seems to use qemu_guest_getrandom which in turn
with patch 1 calls /dev/urandom...
Did I miss something?

> 
> Regards,
> Daniel
> -- 
> |: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
> |: https://libvirt.org         -o-            https://fstop138.berrange.com :|
> |: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

Daniel P. Berrangé May 10, 2019, 4:25 p.m. UTC | #5

On Fri, May 10, 2019 at 12:21:19PM -0400, Michael S. Tsirkin wrote:
> On Fri, May 10, 2019 at 05:16:44PM +0100, Daniel P. Berrangé wrote:
> > On Fri, May 10, 2019 at 12:12:41PM -0400, Michael S. Tsirkin wrote:
> > > On Fri, May 10, 2019 at 03:42:01PM +0200, Laurent Vivier wrote:
> > > > From: Kashyap Chamarthy <[e-mail]>
> > > > 
> > > > When QEMU exposes a VirtIO-RNG device to the guest, that device needs a
> > > > source of entropy, and that source needs to be "non-blocking", like
> > > > `/dev/urandom`.  However, currently QEMU defaults to the problematic
> > > > `/dev/random`, which is "blocking" (as in, it waits until sufficient
> > > > entropy is available).
> > > > 
> > > > Why prefer `/dev/urandom` over `/dev/random`?
> > > > ---------------------------------------------
> > > > 
> > > > The man pages of urandom(4) and random(4) state:
> > > > 
> > > >     "The /dev/random device is a legacy interface which dates back to a
> > > >     time where the cryptographic primitives used in the implementation
> > > >     of /dev/urandom were not widely trusted.  It will return random
> > > >     bytes only within the estimated number of bits of fresh noise in the
> > > >     entropy pool, blocking if necessary.  /dev/random is suitable for
> > > >     applications that need high quality randomness, and can afford
> > > >     indeterminate delays."
> > > > 
> > > > Further, the "Usage" section of the said man pages state:
> > > > 
> > > >     "The /dev/random interface is considered a legacy interface, and
> > > >     /dev/urandom is preferred and sufficient in all use cases, with the
> > > >     exception of applications which require randomness during early boot
> > > >     time; for these applications, getrandom(2) must be used instead,
> > > >     because it will block until the entropy pool is initialized.
> > > 
> > > So how about just using getrandom then?
> > 
> > The 3rd patch in this series addresses that.
> 
> It seems to use qemu_guest_getrandom which in turn
> with patch 1 calls /dev/urandom...
> Did I miss something?

qemu_guest_getrandom will preferentially use the crypto library random
APIs (gnutls, or gcrypt). If both are compiled out that it will use
getrandom() if supported by the C library and current kernel. If that
fails then it will try /dev/urandom if it exists, finally /dev/random. 
On Windows it uses their native crypto API. See this dependant series:

https://lists.gnu.org/archive/html/qemu-devel/2019-05/msg02237.html

Regards,
Daniel

Michael S. Tsirkin May 10, 2019, 4:55 p.m. UTC | #6

On Fri, May 10, 2019 at 05:25:54PM +0100, Daniel P. Berrangé wrote:
> On Fri, May 10, 2019 at 12:21:19PM -0400, Michael S. Tsirkin wrote:
> > On Fri, May 10, 2019 at 05:16:44PM +0100, Daniel P. Berrangé wrote:
> > > On Fri, May 10, 2019 at 12:12:41PM -0400, Michael S. Tsirkin wrote:
> > > > On Fri, May 10, 2019 at 03:42:01PM +0200, Laurent Vivier wrote:
> > > > > From: Kashyap Chamarthy <[e-mail]>
> > > > > 
> > > > > When QEMU exposes a VirtIO-RNG device to the guest, that device needs a
> > > > > source of entropy, and that source needs to be "non-blocking", like
> > > > > `/dev/urandom`.  However, currently QEMU defaults to the problematic
> > > > > `/dev/random`, which is "blocking" (as in, it waits until sufficient
> > > > > entropy is available).
> > > > > 
> > > > > Why prefer `/dev/urandom` over `/dev/random`?
> > > > > ---------------------------------------------
> > > > > 
> > > > > The man pages of urandom(4) and random(4) state:
> > > > > 
> > > > >     "The /dev/random device is a legacy interface which dates back to a
> > > > >     time where the cryptographic primitives used in the implementation
> > > > >     of /dev/urandom were not widely trusted.  It will return random
> > > > >     bytes only within the estimated number of bits of fresh noise in the
> > > > >     entropy pool, blocking if necessary.  /dev/random is suitable for
> > > > >     applications that need high quality randomness, and can afford
> > > > >     indeterminate delays."
> > > > > 
> > > > > Further, the "Usage" section of the said man pages state:
> > > > > 
> > > > >     "The /dev/random interface is considered a legacy interface, and
> > > > >     /dev/urandom is preferred and sufficient in all use cases, with the
> > > > >     exception of applications which require randomness during early boot
> > > > >     time; for these applications, getrandom(2) must be used instead,
> > > > >     because it will block until the entropy pool is initialized.
> > > > 
> > > > So how about just using getrandom then?
> > > 
> > > The 3rd patch in this series addresses that.
> > 
> > It seems to use qemu_guest_getrandom which in turn
> > with patch 1 calls /dev/urandom...
> > Did I miss something?
> 
> qemu_guest_getrandom will preferentially use the crypto library random
> APIs (gnutls, or gcrypt). If both are compiled out that it will use
> getrandom() if supported by the C library and current kernel. If that
> fails then it will try /dev/urandom if it exists, finally /dev/random. 
> On Windows it uses their native crypto API. See this dependant series:
> 
> https://lists.gnu.org/archive/html/qemu-devel/2019-05/msg02237.html
> 
> Regards,
> Daniel

In particular

https://lists.gnu.org/archive/html/qemu-devel/2019-05/msg02238.html


maybe clarify this is just for systems without getrandom then.




> -- 
> |: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
> |: https://libvirt.org         -o-            https://fstop138.berrange.com :|
> |: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

Daniel P. Berrangé May 10, 2019, 5:11 p.m. UTC | #7

On Fri, May 10, 2019 at 12:55:18PM -0400, Michael S. Tsirkin wrote:
> On Fri, May 10, 2019 at 05:25:54PM +0100, Daniel P. Berrangé wrote:
> > On Fri, May 10, 2019 at 12:21:19PM -0400, Michael S. Tsirkin wrote:
> > > On Fri, May 10, 2019 at 05:16:44PM +0100, Daniel P. Berrangé wrote:
> > > > On Fri, May 10, 2019 at 12:12:41PM -0400, Michael S. Tsirkin wrote:
> > > > > On Fri, May 10, 2019 at 03:42:01PM +0200, Laurent Vivier wrote:
> > > > > > From: Kashyap Chamarthy <[e-mail]>
> > > > > > 
> > > > > > When QEMU exposes a VirtIO-RNG device to the guest, that device needs a
> > > > > > source of entropy, and that source needs to be "non-blocking", like
> > > > > > `/dev/urandom`.  However, currently QEMU defaults to the problematic
> > > > > > `/dev/random`, which is "blocking" (as in, it waits until sufficient
> > > > > > entropy is available).
> > > > > > 
> > > > > > Why prefer `/dev/urandom` over `/dev/random`?
> > > > > > ---------------------------------------------
> > > > > > 
> > > > > > The man pages of urandom(4) and random(4) state:
> > > > > > 
> > > > > >     "The /dev/random device is a legacy interface which dates back to a
> > > > > >     time where the cryptographic primitives used in the implementation
> > > > > >     of /dev/urandom were not widely trusted.  It will return random
> > > > > >     bytes only within the estimated number of bits of fresh noise in the
> > > > > >     entropy pool, blocking if necessary.  /dev/random is suitable for
> > > > > >     applications that need high quality randomness, and can afford
> > > > > >     indeterminate delays."
> > > > > > 
> > > > > > Further, the "Usage" section of the said man pages state:
> > > > > > 
> > > > > >     "The /dev/random interface is considered a legacy interface, and
> > > > > >     /dev/urandom is preferred and sufficient in all use cases, with the
> > > > > >     exception of applications which require randomness during early boot
> > > > > >     time; for these applications, getrandom(2) must be used instead,
> > > > > >     because it will block until the entropy pool is initialized.
> > > > > 
> > > > > So how about just using getrandom then?
> > > > 
> > > > The 3rd patch in this series addresses that.
> > > 
> > > It seems to use qemu_guest_getrandom which in turn
> > > with patch 1 calls /dev/urandom...
> > > Did I miss something?
> > 
> > qemu_guest_getrandom will preferentially use the crypto library random
> > APIs (gnutls, or gcrypt). If both are compiled out that it will use
> > getrandom() if supported by the C library and current kernel. If that
> > fails then it will try /dev/urandom if it exists, finally /dev/random. 
> > On Windows it uses their native crypto API. See this dependant series:
> > 
> > https://lists.gnu.org/archive/html/qemu-devel/2019-05/msg02237.html
> 
> In particular
> 
> https://lists.gnu.org/archive/html/qemu-devel/2019-05/msg02238.html
> 
> maybe clarify this is just for systems without getrandom then.

I'm not sure I see what the problem is. That patch is implementing the
fallback behaviour I describe above, with the crypto library preferred,
falling back to getrandom, then /dev/urandom, finally /dev/random.

Regards,
Daniel

Patch

diff --git a/backends/rng-random.c b/backends/rng-random.c
index e2a49b0571d7..eff36ef14084 100644
--- a/backends/rng-random.c
+++ b/backends/rng-random.c
@@ -112,7 +112,7 @@  static void rng_random_init(Object *obj)
                             rng_random_set_filename,
                             NULL);
 
-    s->filename = g_strdup("/dev/random");
+    s->filename = g_strdup("/dev/urandom");
     s->fd = -1;
 }
 
diff --git a/qemu-options.hx b/qemu-options.hx
index 0191ef8b1eb7..4df0ea3aed5c 100644
--- a/qemu-options.hx
+++ b/qemu-options.hx
@@ -4286,7 +4286,7 @@  Creates a random number generator backend which obtains entropy from
 a device on the host. The @option{id} parameter is a unique ID that
 will be used to reference this entropy backend from the @option{virtio-rng}
 device. The @option{filename} parameter specifies which file to obtain
-entropy from and if omitted defaults to @option{/dev/random}.
+entropy from and if omitted defaults to @option{/dev/urandom}.
 
 @item -object rng-egd,id=@var{id},chardev=@var{chardevid}